Where Industry Meets Innovation

  • Contact Us
  • sign in Sign In
  • Sign in with certificate
mit campus

Resources

Search News

  • View All
  • ILP News
  • MIT Research News
  • MIT Sloan Managemet Review
  • Technology Review
  • Startup Exchange

News

November 22, 2017

BROWSE NEWS RESULTS

42 Results | Prev | 1 | 2 | Page 3 | Last | Next
 
StartupExchange
September 26, 2016

Simulation software for super-sized projects

Akselos simulation platform promises to accelerate simulation time to enable complete detailed simulations
In recent decades, simulation software has become widely adopted in infrastructure engineering, helping to reduce development time and identify flaws before they’re fully baked into the design. Typically, engineers use Finite Element Analysis (FEA) computational modeling technologies to simulate systems under many operating conditions in a so-called “virtual prototyping” process. David Knezevic
CTO
Akselos
Yet, FEA is showing its age, especially when it comes to large infrastructure. “With large-scale projects, FEA hits a wall where the computational cost grows very quickly, and this cannot be overcome even with the latest HPC systems,” says David Knezevic, CTO of a MIT simulation technology spinoff called Akselos. “The computational intensity of FEA has led to all sorts of limitations in how it’s used. Engineers tend to either do detailed analyses of small parts of the system or else do a very coarse analysis of the overall system.”

The company’s Akselos simulation platform promises to accelerate simulation time to enable complete detailed simulations from start to finish. “Our software enables detailed analysis of much larger infrastructure than has been possible in the past,” says Knezevic. “When performing a detailed analysis of large infrastructure, you typically obtain 1,000 times or more speed-up compared with conventional FEA simulations.”

The 17-employee company is headquartered in Lausanne, Switzerland, with major offices in Ho Chi Minh City, Vietnam, and Boston, where Knezevic works. Akselos’ customers are building and maintaining infrastructure such as power station gas turbines, wind turbines, offshore oil and gas structures, ships, submarines, and mining infrastructure. Most of these systems are often operating under very extreme conditions. The impressive speed of Akselos and the wide range of industries that could benefit from it are just some of the reasons why the MIT Industrial Liaison Program (ILP) selected the company as one of the first six startups in the new MIT Startup Exchange STEX25 program. This tightly focused spinoff of MIT Startup Exchange is designed to facilitate industry interaction with a select group of the 25 most promising MIT-based startups.

Digging into the Details with Reduced Basis FEA
In 2009, Australian-born Knezevic, a former Oxford Rhodes scholar and Harvard Lecturer, joined Professor Tony Patera’s research group at MIT’s Department of Mechanical Engineering. Working as a postdoc, he pursued simulation research that had been underway there for more than a decade. Like many researchers around the world, Patera was looking for ways to accelerate simulations for large infrastructure in which FEA wasn’t feasible. Patera continues to work on this so-called “Reduced Order Modeling” research today.

The research that Knezevic focused on resulted in what he calls the unique value proposition of Akselos: Reduced Basis FEA, or RB-FEA. RB-FEA builds on top of conventional FEA, but it adds an “acceleration layer” which dramatically reduces solve times.

“The reason that FEA can be slow is that it uses a generic representation of the system, . which is very flexible, but also very computationally expensive,” says Knezevic. By comparison RB-FEA builds up a physics-based dataset for each component in a system. These datasets are reused when a simulation is performed, which leads to RB-FEA’s enormous speed advantage.

Because RB-FEA uses parameterized models, with parameters assigned to each component, you can click on a component in the Akselos GUI and quickly change factors like length, density, stiffness, curvature, Poisson ratio, and other properties. “The parameterized models make it easier to update the model, which is extremely valuable for iterative design,” says Knezevic. “From an early stage in a design process, you can simulate a model hundreds or thousands of times with RB-FEA. In contrast, if you modify something in an FEA model, you have to go through the whole FEA solve all over again, which is typically a thousand times slower than using the Akselos solver.”

Akselos uses a cloud-based infrastructure that “leverages HPC and parallel computing to the fullest extent,” says Knezevic, although he adds that “Akselos’s unique speedup is purely due to the RB-FEA algorithms.” Akselos offers other features one might expect from a modern simulation package, including online model libraries, a drag-and-drop GUI, and a decision support system.

Customers such as oil and gas producers are particularly interested in the substantial CAPEX reductions enabled by Akselos in the upfront design phase. “Industry is increasingly interested in lean design,” says Knezevic. “There’s a strong opinion that a lot of things are over designed. You often see overly conservative models with excessive safety factors, which can be a symptom of incomplete understanding of a system. If you model a system accurately and in detail, you can understand the risks much more precisely and reduce the safety factors while still being extremely safe.”

A leaner design not only reduces construction costs, but also maintenance costs, says Knezevic. “CAPEX reduction up front leads to OPEX reduction over the lifetime of an asset.”

Akselos is especially qualified in the post-deployment maintenance phase. With the speedy, detailed simulations enabled by RB-FEA, engineers can build a highly detailed digital twin of the physical asset, and easily update it as the asset is upgraded or damaged. Alternatively, Akselos designers can build a twin for customers.

“Akselos models are extremely well suited to being updated to account for the reconfiguration of physical assets,” says Knezevic. “Because our models are modular, you can easily replace components or modify parameters, and then re-solve very quickly.”

With Akselos, customers can update the digital twin based on changes made to the physical asset. Better yet, they can make the changes to the digital twin first to understand the impacts on reliability, throughput, and safety before construction. “You can run 1,000 simulations on it for different scenarios, and then you’ll understand if it’s safe or if you need remedial action and maintenance,” says Knezevic.

Sensor Integration
Currently, a key focus for Akselos is to link digital twins to sensor data. Accelerometers, strain gauges, and other sensors are increasingly being deployed on large infrastructure, providing valuable insights into the current state of critical infrastructure. Calibrating digital twins based on this sensor data provides an advanced new approach to structural integrity management.
“With sensor inputs, you have a real-time digital twin that takes into account any physical changes in your asset,” says Knezevic. “You can then run analysis on this updated, calibrated twin to determine risk.”

“If you can detect and react to issues in a predictive and precise way, you can reduce unplanned physical maintenance, which is extremely expensive because you usually have to shut down production,” says Knezevic. “Based on accelerometer input about structural vibration, for example, you can calibrate your digital twin and then run a thousand different simulations on the calibrated model. This allows you to analyze how the infrastructure in its current state would react to different situations, like if a storm hits, or if you want to increase system throughput”

From a Deshpande grant to STEX25
The idea for building a company around RB-FEA emerged in 2011 when Patera, Knezevic, and two of Knezevic’s colleagues -- Phuong Huynh and current Akselos CEO Thomas Leurent decided the RB-FEA technology was ready to be commercialized. They applied for a Deshpande grant at MIT and were funded later that year to continue the commercially oriented research around RB-FEA. This research led to the founding of Akselos in 2012.

“Throughout this process we benefited a great deal from many resources at MIT, including the Deshpande Center, the ILP, and the Venture Mentoring Service,” says Knezevic. “Our VMS mentors acted like a board for us, and gave us great guidance in how to build up a company.”

Even at four years old, Akselos is still benefiting from MIT assistance by being a member of the MIT Startup Exchange program. The company’s selection to the STEX25 should further expand its interactions with MIT’s industrial contacts.

“We’re honored to be part of STEX25,” says Knezevic. “It’s a great way to expand our outreach to companies that are facing challenges with designing, operating, and maintaining complex and critical engineering systems.”

Readers may download and try out Akselos's product for free from this community.akselos.com.About MIT Startup Exchange, STEX25, and MIT’s Industrial Liaison Program (ILP)
MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.

StartupExchange
May 23, 2016

A New Spin on HFT

Domeyard LP leverages diverse expertise sets to push boundaries in high-frequency trading.
The technology world has grown accustomed to stories of successful startups emerging from garages and dorm rooms, but it’s still a rarity in the financial world. Domeyard’s humble dorm-room origin story is only one reason this Boston-based high frequency trading (HFT) firm is turning heads from Wall Street to Silicon Valley and beyond.


Christina Qi
Partner, Domeyard LP


“The support from the finance, technology, and academic communities has been phenomenal. We’ve increased our firm’s value from zero to nine digits, less than a year after launch,” says Domeyard co-founder Christina Qi, who graduated from MIT in 2013.

It helped that the dorm room was at MIT and that other Domeyard founders and early employees came from Harvard and MIT — a union alluded to by the company name. It also helped that other key members of the team have worked at Google, Apple, and Microsoft Research, as well as financial firms such as Goldman Sachs, Virtu, and GETCO.

Qi credits the MIT Startup Exchange for being “a huge help in starting our company.” Last year, Domeyard exhibited at a MIT Startup Exchange R&D conference where the co-founders met a number of clients and partners. One of Domeyard’s lead investors was an MIT Industrial Liaison Program member. “He started his hedge fund from scratch almost four decades ago, after someone decided to take a chance and help him out. Now he’s paying it forward,” says Qi.

In its less than three years of existence, Domeyard has moved on to close multiple rounds of investment. Investors include the founder of one the largest quant funds in the world, the CEO of a global consulting firm, one of the biggest private equity investors in the world, and one of the pioneers of China’s Internet industry, says Qi.

Investors both in the company and in Domeyard’s services are attracted by some of the fastest trading technologies in the world. “To be even a millisecond or a microsecond faster makes a huge difference in many industries, especially in finance,” says Qi. “We have focused on performance, an area in which we could potentially be the best in the world.”



Beyond Flash Boys – the changing face of HFT
The term “high frequency trading” was only coined around 2007. HFT describes a segment of the financial services world that specializes in rapid algorithmic trades using fast servers and networks, marked by high order-to-trade ratios.

The irony of Domeyard’s rapid success is that it was founded only months before the publication of Michael Lewis’ scathing critique of the HFT industry in the best-selling “Flash Boys: A Wall Street Revolt.” The controversial book alleged that trading markets were rigged by HFT traders who specialized in front running orders, a practice that uses extremely high-speed infrastructure and sophisticated algorithms — the sort that Domeyard possesses — to cash in on advance knowledge of pending orders in the market.

HFT’s reputation had already been drawn into question by the “Flash Crash” of May 6, 2010. A $4.1 billion trade on the Chicago Mercantile Exchange initiated with a bank’s automated execution algorithm resulted in the Dow Jones Industrial Average losing over 1000 points before bouncing back to near the previous value, all in a matter of fifteen minutes.

The 2010 Flash Crash, as well as an October 2013 Flash Crash in Singapore, in which $6.9 billion in capitalization vaporized, led to a feverish rethinking of HFT. Exchanges and regulatory bodies began to implement new rules to add some human oversight into HFT, sometimes adding latency in the process. Regulators have also attempted to make trades more transparent in order to expose potentially illegal or destabilizing activities.

“We started Domeyard at a very interesting and eventful time,” says Qi. “The timing was ideal for us because we were forced to confront long-term issues early on, tackling the adversity that large companies face. We looked at the bigger picture of what we’re doing, the impact of our trading strategies, why we’re doing it, and whether it’s good for society. We stayed true to our founding belief that a sustainable business must provide some form of service to the community.”

Domeyard gambles on in-house tech
In addition to rethinking how quantitative trading should be practiced, Domeyard’s management chose an interesting approach to implementing technology. While most software-oriented tech startups take advantage of third party cloud-based services for much of their infrastructure today, Domeyard went for an in-house strategy.

“Unlike other trading firms, we built most of our technology in-house rather than outsourcing to third party vendors,” says Qi. “Most HFT platforms are still new, and not quite as developed as we would like. Going in-house also saves a lot of money. What we have built is of higher caliber and greater value than what is currently on the market. It would cost thousands of dollars per hour on the cloud to spin up the amount of storage and computational cores that we have today.”

Like most HFT-oriented firms, Domeyard takes advantage of physical proximity to glean every last speed advantage. “We co-locate our servers next to the exchanges’ matching engines in locations like New Jersey, enabling us to receive data at a much faster rate,” says Qi. “This is especially true when receiving raw data that you have to clean up and process really fast. Finding signals in data is one feat, but turning them into profitable opportunities takes a tremendous amount of skill and teamwork.”

Increasingly, running an HFT firm requires a lot more than speed. Co-location, fast servers and algorithms, and the latest networking technologies are essential, but “we also need to generate great signals to trade on the marketplace,” says Qi. “Finance is becoming more intertwined with technology. We want to take more of a scientific approach to trading, so it’s not so much based on rumors and gossip, or researching a specific company, but rather about using mathematical models and very deterministic algorithms. We’re eliminating the element of chance. You can’t make tens of thousands of lucky trades in a day.”

Another reason why Domeyard is looking beyond sheer performance is that future improvements in the near-term are likely to be minor. “We are very close to the latency threshold of what’s physically possible,” says Qi. “While we have one of the fastest trading systems around, we are also focusing on more sophisticated strategies. Our edge is in creating smarter strategies without compromising speed.”

Managing the tradeoffs between speed and intelligence is core to Domeyard’s mission. “Running sophisticated algorithms can equate to being slow, so for each situation we have to strike the right balance between speed and complexity,” says Qi. “Our decision depends on intraday market conditions, the markets we’re trading in, and the rules and regulations in each market and jurisdiction. We tailor and revise our strategies every day.”

Domeyard is also notable for being diverse, both in terms of ideas and the people behind them. “Many HFT firms have a common lineage, branching from one of the ten largest firms in the industry,” says Qi. “In contrast, our founding team came from different companies and industries. We had different majors in college. We weren’t fraternity brothers or best friends growing up. Instead, we found each other because of our mutual career interests and goals.”

“When a new hire joins the team, we encourage them to avoid repeating the ideas that their previous firms have implemented,” Qi says. “Just because a large competitor chose a specific solution doesn’t mean that it’s the best solution for us.”

Domeyard has attracted a competitive talent pool thanks to its tech startup atmosphere, academic environment, and flat managerial structure. The latter is still far from the norm in the tech world, and pretty much unknown in financial services. “The goal is that everyone can contribute their ideas openly regardless of age, experience, or background,” Qi explains.

Qi’s main goal for Domeyard is to stay on track and optimize core competencies. “There are a lot of distractions out there, such as new markets we could be trading in, or networking events every other day,” she says. “But we try our best to focus on what we’re doing now. It takes a balance of focus, creativity, and feedback to become the best in the industry.”



MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.
StartupExchange
April 25, 2016

Engineering Tomorrow’s Synthetic Biologists

Natalie Kuldell started BioBuilder to inspire new generations of scientists and engineers with hands-on laboratory experiences.
“If I had only learned science the way it was taught to me in the classroom, I probably never would have become a scientist,” says Natalie Kuldell, a faculty member in MIT’s Department of Biological Engineering. “It was only in high school when I had a chance to work in an investigative lab that I realized how creative and fun science could be.”


Natalie Kuldell
President, Founder, & Executive Director

BioBuilder Educational Foundation



As director of a nonprofit, MIT-spawned startup called the BioBuilder Educational Foundation, Kuldell aims to pass along that spirit of adventure to students. BioBuilder provides a web-based curriculum for synthetic biology aimed primarily at high-school students, and also hosts after-school clubs, teacher training, and other programs. BioBuilder is now taught in more than 40 states and a dozen countries.

“We want to make the emerging field of synthetic biology more accessible,” says Kuldell. “We want students to start thinking about cells as tiny factories and DNA as a programming language that can control the living system.”

As science and engineering research programs and companies compete to attract the brightest young prospects, synthetic biology is hampered by the fact that relatively few students — or adults for that matter — know what it is. Many who do recognize the term often know it in connection with the controversies over genetically modified crops.

Even the experts sometimes disagree on the exact definition of synthetic biology, which attempts to unify biotechnology fields including genetic engineering, molecular and evolutionary biology, and even computer engineering. According to the MIT Synthetic Biology Center, the goal of synthetic biology is “to make the construction of novel biological systems into a practical and useful engineering discipline.”

BioBuilder’s website points to a variety of real-world applications beyond food engineering, including the development of biofuels, anti-malarial drugs, biodegradable adhesives, and less toxic cancer treatments. “The hope is that synthetic biology can turn biotech into something reliable, robust, and scalable in the same way that engineering turned physics and chemistry into technologies we can all rely on,” says Kuldell.



That many high school students have never heard of synthetic biology is not surprising. Between their lack of funding and lack of time, high school science classes struggle to squeeze in the basics of science, let alone dig into the most recent biotech breakthroughs. Science teachers must often resort to time-efficient lectures rather than hands-on investigations and lab work that more effectively engage students.

Keeping Synthetic Biology Real
Kuldell conceived the idea for BioBuilder when she realized much of the curriculum she had developed over a dozen years of teaching synthetic biology to MIT students could be adapted to high school students. The BioBuilder program differs from most K12 science curricula in that it is regularly updated with the latest research underway at MIT and elsewhere.

“I’ve always tried to base my teaching around authentic research questions, which leads to collaborative and team-based learning, and lasting moments of engagement,” says Kuldell. “BioBuilder takes these real research questions and converts them into teachable modules framed with engineering challenges.”

Another major goal for the program is to make the education as hands-on as possible in order to “engage students as creative and critical thinkers,” says Kuldell. The multimedia-rich BioBuilder curriculum and hands-on lab kits lead to an online portal that lets students share what they’ve learned.

Activities include exploring bacterial photography, evaluating identical DNA programs in different types of bacterial strains, and varying the protein output from a series of genetic devices. In one BioBuilder activity, students try to generate bacteria that smell like bananas.

“We ask the students to design a genetic program that will regulate the output of cells to only smell pleasant during a particular phase of growth,” says Kuldell. “It may seem silly, but the scent industry is enormous. Scent has been an under-utilized reporter for cellular behavior, and the ability to control cell outputs is key to any biotechnology.”

BioBuilder keeps the teaching relevant by showing the potential for synthetic biology to solve real-world problems like hunger, climate change, and disease. The curriculum also covers biosafety, bioethical issues, and the debate over genetically modified foods.

BioBuilder offers more of an engineering focus than is typically found in high school biology classes. “I am a scientist by training, and when I first came to MIT I did not fully appreciate the work engineers do,” says Kuldell. “I came to realize how engineers can apply what we learn through science in order to meet real world needs. I saw how effectively these engineering challenges could be used to engage students and teach them the science, as well as the limitations of scientific understanding.”

Synthetic biology takes “an engineer’s eye and applies to it biology,” says Kuldell. “In life sciences, we need to move beyond ad hoc endeavors when we are putting pieces of DNA together and develop protocols and standards so they can be assembled in a reliable way. Through standardization and shared databases, we may be able to lower the barriers of entry of doing biotech to the point where everybody can do it.”

MIT Venture Mentoring Service helps chart trajectory
The heart of the BioBuilder project, which focuses on investigative curriculum for high schools, started with a multiyear grant from the National Science Foundation. Halfway through that funding cycle, the NSF requested details on how the project could be made self-sustaining, and Kuldell sought out MIT’s Venture Mentoring Service for advice.

“Working with the Venture Mentoring Service, it became clear that by establishing the project as an independent nonprofit, we could build capacity and expand into the broader community,” says Kuldell.

Four years ago, the VMS helped Kuldell plan the BioBuilder Educational Foundation. Since then, the Foundation has been sustained with individual contributions, as well as partnerships with companies that sell BioBuilder’s lab kits, and through additional grants.

Assistance from the VMS has been extended post launch to help Kuldell expand the program. “They were critically important in helping me think about how to scale the work so I didn’t have to be everywhere,” says Kuldell. “The VMS showed me how teachers could become ambassadors for the program. I hadn’t anticipated this wonderful community of teachers that has supported BioBuilder. Teachers run teacher training workshops all over the country, and promote BioBuilder at professional meetings. They have led the way in adapting the curriculum for middle schools.”

O’Reilly Media recently published a BioBuilder textbook, now available on Amazon, which has helped teachers who run into “institutional resistance” from schools and school districts, says Kuldell. “The textbook helps bring synthetic biology into classrooms as a formal curriculum. We are also starting an after school club for schools that aren’t ready to bring it into the formal classroom but would like to try it out.”

BioBuilder plans to launch a dedicated BioBuilder teaching and learning space in Kendall Square with Lab Central, which provides lab space and an entrepreneurial environment for biotech startups. The partnership is one of many the foundation has forged with industry — mostly Boston-area firms — in order to expand BioBuilder’s reach.

“The Massachusetts Life Science Center provided equipment grants to our teachers, and through that I’ve developed a relationship with VWR’s Wards Science, which has been distributing our lab materials,” says Kuldell. “Now New England Biolabs wants to provide reagents to our clubs. Biotech businesses realize it’s important to bring community and public access into the biotech world.”

Just as BioBuilder has given high school students the confidence to envision careers in synthetic biology, it has also helped Kuldell gain confidence in her own abilities. “I could not have imagined myself running a public benefit organization,” she says. “There’s something about being at MIT that makes you very brave and makes you believe that you can do the work that you think is important.”



MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.
StartupExchange
March 21, 2016

Turning Up the Heat Exchange

DropWise Technologies’ coatings promise major energy savings for power plants and other vital industrial uses.
In a way, we’re still living in the Steam Age, and grappling with the limitations of steam power. More than 85% of the world’s electrical power comes from steam power plants, according to Adam Paxson, chief executive officer of DropWise Technologies. After driving the plant’s turbines, steam is condensed back to water in a heat exchanger, whose metal tubes are filled with running cold water. This condensation also creates a vacuum that helps to pull steam through the turbines. But condensation itself is not so efficient, because thick blankets of water build up on the condenser’s metal tubes and block the flow of heat.


Adam Paxson
DropWise Technologies CEO


DropWise Technologies, a startup based on research from two MIT labs, has developed ultra-thin, ultra-effective coatings to break down the blankets of water and bring major savings in energy and water use for these power plants. Just as crucially, the water-repellant coatings could fight climate change by cutting carbon dioxide production.

“Even a small incremental improvement across heat exchange efficiency can have enormous global impact, and with these coatings, we’re talking about heat exchanger coefficients going up by a factor of seven,” says Paxson. “Implementing this coating in a typical large-scale power plant would offset the amount of CO2 equivalent to a few thousand cars. And for the first time, retrofitting existing power plants can be done in a way that is economically and operationally viable, with just a truck outfitted with a small amount of chemicals and deposition equipment.”

Beyond power plants, almost every major industrial process makes heavy use of heat exchangers, he points out. DropWise also targets chemical processing, desalination, turbines, power components and other industrial applications that have not been addressed successfully by other coating technologies.

Coat dependent

Heat exchangers transfer heat from one fluid to another through a very thin sheet of metal, “which inevitably gets fouled or corroded by some form of material — whether it’s thick films of condensing steam that build up on these surfaces, or biofilms or corrosion,” Paxson explains. “Even a very thin layer, a micron thick, makes a significant impact on performance.”



Working on his PhD in the lab of Kripa Varanasi, associate professor of mechanical engineering, Paxson realized the need for a simple, highly manufacturable solution to the heat exchange problem, with coatings that were very thin, highly robust, and easily applied to large surfaces.

Paxson and Varanasi created advanced coating designs and collaborated with Karen Gleason, MIT associate provost and professor of chemical engineering. “We combined the coating processing and materials that Professor Gleason had been developing with the application expertise and testbeds in the Varanasi lab,” Paxson says. “Right away we started getting some incredible results in terms of both performance and durability.”



David Borrelli
DropWise Technologies CTO



The coatings are generated by an initiated chemical vapor deposition (iCVD) technique created by the Gleason lab, says David Borrelli, formerly a PhD student in the lab and now DropWise’s chief technology officer. The iCVD technique flows gases across hot filaments to graft ultra-thin polymers to metal surfaces, while maintaining the metal surfaces at room temperature.

This process creates an extremely thin film, about one two-thousandth the thickness of a piece of paper. “It’s very difficult by any finishing process to get a thin enough coating that the coating itself doesn’t inhibit your heat transfer,” says Borrelli. “That’s one of the key benefits in our processes.” Another is that the coating can be applied across large expanses of metal or other surfaces.

Two aspects of the technology impart durability, a crucial requirement. The polymers that are being deposited bond tightly with the metal oxides on the surface of the heat exchangers, and the polymers create a cross-linked overlying layer that protects against chemical reactions driven by the steam.

Paxson built a testbed in the Varanasi lab, essentially a miniature power plant that simulates the temperatures and pressures inside an actual operating power plant. Accelerated testing at higher temperatures has shown no degradation of coating performance over more than three years, and DropWise is running even longer-term durability tests.

“The coating is thick enough to impart a huge amount of durability but also thin enough that it doesn’t have any negative impact on the performance of the heat exchanger,” Paxson says. “This is the first coating technology that can meet all of those technical requirements and is economical enough that a plant can quickly recover the cost of applying the coating.”

Steaming ahead

Seeing a massive commercial opportunity for these coatings, as well as important environmental advantages, Paxson and Borrelli founded DropWise with their professors in 2014.

MIT’s entrepreneurial culture and resources have been particularly helpful in the difficult-to-enter market for advanced materials, Paxson says. With an office in north Cambridge, “we benefit from having the intense brainpower, both the professors and the potential employee pool in the MIT ecosystem, within a ten-minute walk,” he says. Additionally, many promising applications for the coatings have come in through the MIT community.


Annica Blake
DropWise Technologies COO



Sorting out the best early applications is a crucial business question for DropWise, because there are so many potential uses for the coatings. “You can increase heat transfer, or reduce fouling or corrosion of components, and the iCVD process is so flexible that it can be applied in large scales very economically or in small components that have very complex shapes,” notes Annica Blake, chief operating officer. “The choice is, what do we focus on?”

“The end goal for us has always been to get to power plants, because of the impact on the environment and the commercial viability of that market,” Blake says. “But we realize that will take a number of years, because it is a conservative industry and it operates at a large scale. So we’re looking for smaller stepping-stone applications that we can quickly address and bring to market.”

DropWise is running pilots for several such applications with commercial partners today and is optimistic that some uses for its coating technology will start to bring in revenue in the short term.

In September, the company announced a joint-development partnership with Henkel Corporation, a global supplier of coatings and related products. Henkel gives the company considerably more credibility with potential customers, Paxson points out, since the DropWise technology has been rigorously tested by the firm that invented many of the standard tests used by industry globally.

Environmental benefits remain a key driver for DropWise. “Applying this coating to power plant steam condensers can significantly increase the efficiency of the condensers, which increases the efficiency of the power generation cycle, which in turn lowers emissions and lowers water usage,” says Borrelli. “So applying this globally could have a huge impact on CO2 emissions.”

In fact, says Paxson, adopting the coatings in power plants could lower CO2 emissions more than all the solar energy equipment installed worldwide in a year. “The personal motivation behind this startup was the realization that with a handful of smart people, we can have more environmental impact than entire global energy industries,” he says.



MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.
StartupExchange
February 22, 2016

Disrupting Drug Discovery

twoXAR’s analytics combine radically different data sets to match up diseases with potential drug candidates.
“We’re looking to build a next-generation biopharmaceutical company that brings a data science-first approach to drug discovery,” says Andrew M. Radin, co-founder and chief business officer of twoXAR. “This is where we can have the biggest impact in reducing the time and costs associated with finding more effective medicines across many complex diseases, especially in rare diseases where there is a lack of investment.”


Andrew M. Radin
CBO & Co-Founder

twoXAR



A startup in Palo Alto, California, twoXAR integrates and analyzes massive biological, chemical and clinical data sets to prioritize drug compounds as candidates for treating specific illnesses. “We’re computer scientists solving a biology problem, as opposed to biologists trying to solve a computer science problem,” says Radin. “We have no wet labs and conduct no animal studies.”

The company was founded by two Andrew Radins in 2014 — Andrew M. Radin, who graduated from MIT’s Sloan School that year, and Andrew A. Radin (no relation), a data scientist who has worked as chief technical officer for several startup firms.

twoXAR’s roots go back to a graduate school class Andrew A. took in bioinformatics at Stanford University. Given a homework assignment to extract findings from sets of biomedical data, he created a big-data algorithm to study type 2 diabetes, and produced a surprisingly successful model for predicting which drugs might work for the disease.

That success eventually led to the formation of twoXAR, which has built a commercial computational drug discovery platform based on the algorithm Andrew A. developed for that initial classroom project. Since then, they have worked painstakingly to evolve the technology and validate the results it generates.

In Parkinson’s disease, for example, twoXAR’s platform sifted through an extremely broad collection of data from pre-clinical and clinical research, along with a library of drug compounds, and produced a list of intriguing drug candidates.



At the time, the company didn’t have the expertise in Parkinson’s to evaluate these results. However, it found that one of the top candidates was being studied in the lab of Tim Collier, a leading Parkinson’s researcher at Michigan State University. After examining the twoXAR results in detail, Collier and his colleagues agreed to begin an ongoing collaboration with twoXAR. The Collier lab is now running animal studies to examine the efficacy of some of the compounds the company has identified.

twoXAR also works with scientists at the University of Chicago and Mount Sinai Hospital in New York City. In these academic collaborations, researchers provide the company with data around a disease they’ve been studying. twoXAR puts that data into its system, along with related data sets from various public and private sources, and generates promising candidates. “Basically, at this stage we’ll file indication patents for these candidates, our partners will run animal studies on them, and we’ll share the upside from any discoveries we make together,” Radin says.

Additionally, the startup is engaging with a number of biopharmaceutical firms in drug discovery collaborations. “Each one of those conversations looks very different, depending on the disease in which they’re interested and the stage they’re at in adopting large-scale data sciences,” Radin comments. “We’re not looking for folks to convince; we want to find companies that recognize that the analysis of large data is the path to the future.”

twoXAR offers to help its partners identify new candidates and new targets for a specific disease, prioritize existing candidates for a disease, or validate existing compounds repurposed for another disease.

Bringing data science to bear

From a drug developer’s perspective, “you can think of our platform as similar to high-throughput screening of drug candidates, but much faster, with a much broader set of drugs and data,” says Radin. “Because, when we look at biological data, chemical data and clinical data, these are radically different data sets. Any one of those data sets independently might not provide enough information, but when we look at the overlaps between them, and we start to see the same signal out of all that noise, that’s a strong indication that a drug might treat that disease.”

The twoXAR drug discovery software platform works in a four-step process. The first step is collecting biological data from sources such as gene expression microarrays, chemical data such as molecular structural information, and clinical data to see if the drug might be protective against a similar disease. The data collection also draws on libraries of molecular drug candidates, often using one library with more than 25,000 compounds.

Next, the platform takes that data and plugs it into a network model of the illness. Third, its proprietary algorithms identify relevant features. Finally, these features are plugged into a machine-learning algorithm, which produces a list of the compounds ranked on the probability that they can effectively treat the disease.

Compounds that post the highest scores but are neither known treatments nor under study are particularly interesting candidates for study. “Some may lead to new mechanisms of action, which are what actually cures these diseases and doesn’t just treat the symptoms,” Radin says.

Although current drugs for Parkinson’s, for instance, only treat symptoms, “we can tune our algorithm to focus on things that are potentially neuroprotective — stopping or reversing the progression of disease,” he says. “We’ve filed for patents for several candidates that we’re looking to move forward in studies.”

So far, twoXAR has run analyses for compounds addressing more than 20 diseases. And as the firm broadens the scope of illnesses under study, it also steadily widens the data sets on which it draws. “The more data we have, the better,” Radin notes. “As this universe expands and more people are willing to share their data from public and private sources, we are able to make even more robust predictions.”

Not all of these data sets are equally trustworthy, but “we have algorithms to determine which data sets are noisy and which aren’t, and we throw out the ones that are not relevant,” Radin says. “So it’s not a huge problem for us.”

Connecting and competing

MIT connections have been key for twoXAR’s launch. In the company’s earliest days, Andrew M. drew on conversations with MIT alumni in senior management positions in biopharma companies for guidance on the pharmaceutical industry. He now works with the Industrial Liaison Program, the Martin Trust Center for MIT Entrepreneurship, and former professors at Sloan to find partners and hone business strategies.

Given the rapid changes and deep competition in drug discovery, entrants such as twoXAR have their work cut out for them. “Our competition is basically any pharmaceutical company that’s discovering new drugs,” Radin says. “But really, our biggest challenge is changing the perception of big data-driven approaches within the industry. It is still early days and a lot of skepticism remains about this radically new approach to drug discovery.”

“We're just doing what scientific researchers have always done,” he comments. “But advances in statistical methods, our proprietary algorithms, and secure cloud computing allow us to do it orders of magnitude faster across disease areas where there are real needs for new medicines.”



MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.
StartupExchange
February 1, 2016

From MIT to Startup in Seven Steps

Zaiput Flow Technologies is bringing innovative tools for continuous flow chemistry to market.
As a research associate at the at the MIT Department of Chemistry’s Jensen Lab, Andrea Adamo came up with an idea for a device that separates liquids in the context of continuous flow chemistry. Adamo’s liquid-liquid separator, which uses membrane-based separation and a novel on-board pressure controller, greatly simplifies the liquid separation processes used in pharmaceutical research. In 2013, Adamo joined with Harvard biochemist Jennifer Baltz to launch a Cambridge, MA-based company to commercialize the technology.


Andrea Adamo
Founder & CEO

Zaiput Flow Technologies


Zaiput Flow Technologies is thriving in the research market, and is now readying a larger version of the separator. In 2015, the company was announced as a winner of the Galactic Grant Competition along with Nanobiosym. The grant will help Zaiput develop a version of the separator that can be tested at the International Space Station to research the effects of zero gravity on flow chemistry.

The journey from inspiration to startup to outer space did not happen by accident. Here are seven steps in Zaiput’s journey from lab to marketplace.

The Inspiration
As a child in Italy, Andrea and his brother dreamed of launching their own company. But first they had to find the perfect name. The boys spent hours dreaming up names before they found their winner.

“Zaiput doesn’t mean anything, but we liked the sound of it,” says Adamo. “When I finally had a chance to start a company, I thought why not? Some people love it, some hate it, but people remember it.”

The Foundation
Despite his entrepreneurial stirrings, it was science that captivated young Andrea’s attention. After earning a Ph.D. in Fluid Mechanics from the University Federico II of Naples, he came to MIT 15 years ago on a Fulbright scholarship. At MIT, he added another Masters in Science degree and began working as a research associate in the laboratory of Klavs Jensen, Director of MIT’s Chemical Engineering Department.



At Jensen’s lab, Adamo worked on research such as molecular compound delivery to cells, lab-on-a-chip applications, and microfluidic-based detection systems. He was particularly interested in creating devices for continuous flow chemistry, a process in which chemical reactions are achieved by mixing them together as fluids moving through tubes.

“With flow chemistry, we try to have things react as they flow as opposed to reacting in batch in a vessel,” explains Adamo. “The idea has been around for 50 years or more — what is new is applying the approach to a smaller scale, as well as adding high-added-value molecules. Flow can deliver robustness and quality and expands the parameter space that you can use in reactions.”

Pharmaceutical companies have recently begun investing in plant upgrades with flow chemistry and other “continuous manufacturing” technologies. Such designs can cut costs and improve quality by reducing the number of steps, devices, and locations required for drug manufacturing. A Wall Street Journal story last year quoted Bernhardt Trout, director of the Novartis-MIT Center for Continuous Manufacturing, as saying continuous manufacturing upgrades can save 30 percent or more in operating costs.

Big pharma’s attraction to flow chemistry reignited Adamo’s entrepreneurial ambitions, and he began to brainstorm how the technology could be applied to the problem of liquid-liquid separation. Current separators have numerous drawbacks, says Adamo.

“In typical separators, you use a funnel shaped container in which liquids separate by gravity,” he says. “You mix things up and wait for them to separate. The problem is that some emulsions take forever to separate, and there’s a cost problem in that the containers can be quite large, and the compounds can be very expensive.”

Adamo was fascinated by a development of Jensen’s in which surface forces were used in a flow chemistry system to separate liquids rather than using gravity. With this technique, different liquids flowing through the same porous membrane react differently.

“It’s kind of like a nonstick pan, which causes different reactions with different kinds of liquids,” says Adamo. “If you pour oil onto it, it will spread out. If you pour water, it beads up.”

The Eureka
Adamo pursued a membrane-based design, but faced an obstacle between concept and workable prototype. “In order to have complete separation, you have to precisely control the pressure on either side of the membrane,” says Adamo.

Adamo soon came up with the solution: a high-precision differential pressure controller that sits on top of the device and ensures the membrane always has the correct separating condition. Adamo’s liquid-liquid-separator had the advantage of using a continuous process like an assembly line.

“The shaking and mixing is done by the flow itself, so you don’t have to wait,” says Adamo. “You can also build a cascading system so you have reaction, separation, reaction, and so on. It saves the chemist the time required to play with different pressures when working on complex applications. You assemble molecules and eventually end up with a product such as a drug or a perfume.”

Jensen, a pioneer in flow chemistry, whose recent spinoffs include SQZ Biotech, encouraged Adamo to explore the technology’s business potential. Adamo was particularly inspired when he showed an early prototype to Tim Jamison, the head of MIT’s Chemistry Department, and now a member of Zaiput’s advisory board. “He told me the prototype was really useful, and that I should have considered selling it,” says Adamo.

The Entrepreneurial Education
Over the years, Adamo had drawn inspiration from the entrepreneurial atmosphere at MIT and the surrounding startup community. Now he had to learn how to become an entrepreneur himself.

Adamo had already received Deshpande funding for another business idea he had formulated at MIT, which “provided a tremendous opportunity to interact with mentors,” he says. “At all the networking events, classes, and talks at MIT, I discovered the beauty of entrepreneurship. I learned how to start thinking about a business, how to identify opportunities.”

To prepare for the launch of Zaiput, Adamo took several business school classes, and attended industry conferences. “At one chemistry conference, I realized that my view of the product’s potential was quite narrow,” says Adamo. “So I looked for more connections and inputs to identify other opportunities.”

Adamo also signed up for the MIT Venture Mentoring Service. “VMS gave me some great advice, and helped me structure the business,” says Adamo. “When you go from theory to practice, there are so many aspects to look at. I also had a great interaction with the MIT Technology Licensing Office.”

The Launch
Adamo’s entrepreneurial education was a big help, but it was no replacement for experience. He wisely chose a co-founder in Baltz who contributed experience in both biochemistry and biotech startups. With the help of MIT institutions like VMS, Adamo (CEO) and Baltz (COO) lined up funding to launch Zaiput in 2012.

Zaiput opened an office in Cambridge, put the finishing touches on the liquid-liquid separator, called SEP10, and unveiled it at the 2013 Flow Chemistry Congress in Boston.

The company quickly found interested customers. After lining up a manufacturer and shipping product, Zaiput added a line of more traditional bench gear: tunable, clog-resistant back pressure regulators.

Zaiput’s location has been a big benefit. “Cambridge is the place you want to be,” Adamo says. “Within a few blocks we have all the greatest pharma companies, as well as the future leaders. We can quickly get connected with opinion leaders and buyers alike, so we don’t have to travel so much.”

Adamo, who is still a part-time MIT research associate, continues to draw on MIT institutions for guidance, and he is also active at the Novartis-MIT Center for Continuous Manufacturing.

The Expansion
Adamo’s conversations with customers, mentors, and colleagues led him to realize that scalability was critical. Zaiput is now preparing a larger SEP200 version of the separator designed for a typical flow rate of 200 ml/min, compared to 0-12 ml/min for the current device, and has begun designing an even larger one. The new devices support larger-scale research efforts, and even small-scale manufacturing. Adamo envisions variously sized models for use in clinical trials, where pharma companies can scale up to larger volumes at each stage.

Zaiput is planning to build devices designed “in arrays for complex extraction cases,” says Adamo. When you separate some liquids, such as wine and oil, 10 percent of the alcohol goes into the oil. “If you want to separate out more, you would need to repeat in a cascading system,” he adds. “Today, this process is done with columns 30 meters high. We think we can do it with boxes full of devices.”

Zaiput is also exploring entirely new applications in the food industry and in water remediation. “You could use a version of this to clean up a gasoline spill in a lake,” says Adamo. “You could probably even recover the gasoline and reuse it.”

Although Adamo says the potential for separation technology is “boundless,” he is in no rush to jump into new industries. “As a startup, we are trying to stay focused.”

The Liftoff
Zaiput is now looking for space to set up its own research lab, and there are plans to launch a consulting service. The company recently partnered with Snapdragon Chemistry, another MIT spinoff, which should help it acquire “state of the art flow chemistry knowledge,” says Adamo.

Adamo is thrilled Zaiput won the Galactic Grant earlier this year from the Massachusetts Life Sciences Center (MLSC) and the Center for Advancement of Science in Space (CASIS). The grant means more than just good PR. “The zero gravity research at the space station may provide important results to improve our products,” says Adamo. “This could also be one of the first steps to enable drug making in space, which you would need for long manned missions.”

As for Adamo’s brother who helped come up with the Zaiput name, he’s now playing a peripheral role in the company, as well. He married Andrea’s co-founder and COO, who is now Jennifer Baltz Adamo.



MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.
StartupExchange
December 21, 2015

Reimagining the Role of the Car

Emilio Frazzoli leads autonomous vehicle research efforts aimed at enabling driverless car-sharing fleets.
MIT Professor or Aeronautics and Astronautics Emilio Frazzoli is one of the world’s experts on autonomous vehicles, both in the air on the ground. So when we sat down to talk to him about his autonomous car technology firm, nuTonomy, we expected to hear a lot about computer vision systems, safety algorithms, real-time operating systems, and the latest sensor technologies. But Frazzoli wanted to talk about something much bigger: reimagining the role of the car.


Emilio Frazzoli
MIT Professor of Aeronautics & Astronautics

nuTonomy CTO


Frazzoli believes the self-driving car will fundamentally change the role of the automobile from an extension of our bodies to a shared resource, from mechanical horse to motorized appliance. “One way we can think of autonomous cars is as appliances that happen to move from place to place, perhaps something like an elevator,” says Frazzoli, nuTonomy’s CTO.

While Frazzoli concedes that this conceptual switch “could be controversial,” and that it “will not help selling the car as a product,” he’s not particularly concerned. Frazzoli believes that the first wave of autonomous vehicles will not be sold as personal vehicles, but as services. In particular, nuTonomy is going after the ride-sharing and fleet management industries.

“It could be decades before we will be able to walk into a car dealership and buy an autonomous car and have it drive us home,” says Frazzoli, who launched nuTonomy in 2013 with CEO Karl Iagnemma, director of MIT’s Robotic Mobility Group. “But much sooner than that we will see autonomous vehicles offered as mobility-on-demand services running in well-defined locations during the day under good weather conditions. We don’t want to develop the vehicle itself, but rather provide software and system design, including the selection of sensors and actuators. We hope to eventually provide services.”

NuTonomy started as a consulting business for Tier 1 suppliers and automotive manufacturers like Jaguar/Land Rover, helping them add more autonomous functions to their human-controlled vehicles. More recently, Frazzoli and Iagnemma have set their ambitions higher, and now nuTonomy’s main focus is on developing software for fully autonomous cars aimed at the service industry. Angel investors have anted up, and now the company is looking for more substantial investments.



While Frazzoli says that both semi-autonomous and autonomous approaches will succeed in the coming decades, Nutonomy is betting on a fully autonomous model over the long run. NuTonomy’s vision is closer to that of Google’s self-driving car than the automobile industry’s numerous assisted driving projects. In part this is due to the difficulty in handoffs between human and computer drivers, but it also stems from nuTonomy’s focus on car sharing and fleet services.

“For the service model, we favor a clear demarcation between the authority of the automation vs. the human,” says Frazzoli. “When one is in charge it should not rely on the other taking over and vice versa.”

NuTonomy’s software provides a unique approach to autonomous decision-making. “Our advantage is that our cars can automatically and systematically satisfy all the rules of the road,” says Frazzoli. “Back when we were doing the DARPA Urban Challenge, we had to hand-code all the different options for all the different things that could happen on the road. This is very hard to do by hand and even more painful to debug.”

Having “vowed to never do that again,” Frazzoli and Iagnemma, came up with a more systematic solution that not only saved countless hours of coding, but resulted in a more reliable system. “We provide the software with a list of rules, and then the car automatically satisfies them without the need for hand-coding.”

The software is also flexible enough to know when a traffic rule should be broken. “We teach our cars to use judgment to violate some of the rules when it is both safe and necessary, like driving around a double-parked car,” says Frazzoli. “But there is a danger in trying to humanize the car. Autonomous cars are not meant to replicate the human. That’s tantamount to saying humans are the perfect paradigm for driving, and they’re not.”

Transforming Urban Landscapes with Mobility as a Service
After spending much of his career researching Unmanned Aerial Vehicles, Frazzoli came to MIT and began working on autonomous car research. He was a member of MIT’s 4th-ranked team at the 2007 DARPA Urban Challenge, and later joined the MIT SMART (Singapore MIT Alliance for Research and Technology) program on the Future of Urban Mobility. Frazzoli is now the Lead Principal Investigator at SMART’s autonomous car project, which recently demonstrated the world’s first public autonomous vehicle pilot in Singapore.

The experience of envisioning Singapore’s transportation future helped Frazzoli realize that “this technology has a big potential for a very profound impact on our lives.” Typically, says Frazzoli, people talk about benefits such as safety, as well as the reduction of congestion and smog due to increase driving efficiency. In the case of fully autonomous technology, added benefits include improved accessibility for those who can’t drive, and the greater convenience and productivity of being able to do other things in the car instead of driving.

“Yet these benefits only improve on the status quo,” says Frazzoli. “I realized that the killer app for autonomous cars is car sharing. Cars cost a big chunk of our disposable income, and yet they sit idle 95 percent of the time, often using up expensive real estate. We pay for the privilege of not using the car. If the car can drive itself why leave it parked in the garage? Let it go pick up and drive somebody else, take your children to school or take your spouse to their job, or share it publicly.”

The problem with current car-sharing services is either the lack of availability of a car or a parking lot, says Frazzoli. “ZipCar requires you to return car to same location you picked it up, which limits flexibility,” he says. “The service is experimenting with something like Car2Go’s one-way service, but it doesn’t solve the problem of having to park the car. Self-driving cars eliminate all these problems. We envision a model in which whenever I need a car, I can book it on my phone, and it will pick me up. When I get to my destination I don’t have to worry about parking or refueling.”

Autonomous car sharing could also change the way people think about cars, says Frazzoli. “This could actually be liberating for car guys like myself,” he says. “Like many people I like driving fancy sporty cars, but because I am raising three children, I bought something more practical. What if I could share an autonomous minivan for commodity transportation, taking kids to soccer practice and grocery shopping, and then still have a fancy car to drive on weekends? Many of us can get by without a second car, and can buy the cars they like rather than the cars they need.”

If this vision of shared driving doesn’t do much to cheer car manufacturers, it’s certainly intriguing to fleet operators. “When you compare the cost of a driving service, including the cost of the car itself, with the standard way of providing the service with manned vehicles, it’s very compelling,” says Frazzoli. “Say a taxi driver makes $50,000 a year, and I need to cover three shifts, so it’s now $150,000 a year. We are amortizing the higher expense of a self-driving car over the life of the service and the car.”

At the SMART project, Frazzoli studied the potential impact of shared autonomous vehicles in Singapore. He found that about 300,000 shared autonomous vehicles could provide the same level of mobility available now with 800,000 passenger vehicles.

“You could sell those half million cars to another country, and reclaim the space of a million and a half new parking spaces, which can be given back to people for entertainment, residences, businesses, and parks,” says Frazzoli. “This is very important in a geographically constrained place like Singapore. A lot of the space in big cities is devoted to parking lots, so there’s a big potential in urban development. I grew up in Rome, which is a beautiful city, but the effect is spoiled by double- and triple-parked cars. Our beautiful hidden monuments are hidden behind layers of metal and rubber.”

The path to nuTonomy’s mobility as a service vision will take time. The company is planning a staged rollout, starting with very controlled environments, such as daytime shuttles at corporate campuses, before moving out to car sharing and taxi services over prescribed routes on public roads.

“In the short term, we are looking for the sweet spots where we can offer a service that does not require the car to face bad weather or other difficulties,” says Frazzoli. “In the long run, however, this has the opportunity to change the way people think about cars forever.”



MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.
StartupExchange
November 23, 2015

Defending the Brain

Oxalys Pharmaceuticals discovers drugs with robust cell-culture models of neurodegenerative diseases.
There are no drugs that stop the progression of Huntington’s disease, a rare and fatal genetic condition that afflicts about 30,000 people in the United States. Ditto for Parkinson’s disease, affecting about a million people in this country, and Alzheimer’s disease, now seen in 5 million people in the U.S. and on a massive upswing.


Katharine Sepp
Co-founder & CEO

Oxalys Pharmaceuticals


Current drugs for these neurodegenerative diseases, which can help to reduce some symptoms, work by tweaking the signaling that occurs between neurons. “The problem with those therapies is that as the brain ages the neurons become more and more unhealthy, so this approach is like trying to optimize the running of a broken machine through greasing parts, but what is really necessary is replacement or restoration of the broken components,” says Katharine Sepp, chief executive officer and co-founder of Oxalys Pharmaceuticals. “What you want to do is create a healthier brain overall.”

Oxalys aims to do just that, using an advanced cell-culture platform developed at MIT for finding compounds that slow or stop the advance of neurodegeneration.

The Toronto-based company has received an orphan drug designation from the U.S. Food and Drug Administration (FDA) for a drug candidate for Huntington’s disease, and expects to launch a clinical trial for its drug within two years.

“There’s a tremendous unmet medical need for Huntington’s disease,” Sepp points out. “Huntington’s is a complex neurological disorder that involves psychiatric changes, cognitive decline and changes in motor coordination. There’s only one FDA-approved drug, which lessens motor symptoms but does nothing to improve the cognitive decline or the psychiatric changes. Patients really struggle. It’s very devastating. We need to find something that will treat the disease in an acceptable way.”

Models of Degeneration
Scientists have made major progress in understanding the genetic mechanisms that drive diseases of the aging brain such as Huntington’s, but creating in vitro models of the conditions for testing drug candidates has been challenging.



One earlier “cell-free” method of drug discovery searched for compounds that could break up the aggregations of disease proteins that appear in the brain cells of people with these conditions. However, the compounds these studies identified often were toxic to cells, and there’s an ongoing scientific debate about the roles the aggregates play in the disease. Another approach created neurons that could grow indefinitely in culture, but it didn’t reflect the normal aging process in neurons.

Working in the lab of Troy Littleton, MIT professor of biology, at the Picower Institute for Learning and Memory, Sepp and Oxalys co-founder Joost Schulte developed a very different cell-culture technique, which expresses a high proportion of the human Huntington’s disease gene within Drosophila fruit-fly neurons.


Joost Schulte
Co-founder & CSO
Oxalys Pharmaceuticals


In these cultures, “you can tell clear differences between the freshly prepared neurons expressing the human disease gene versus neurons that don’t carry the human disease gene,” Sepp says. Neurons that express the human disease gene display axons and dendrites that are shorter and more likely to branch, as well as aggregations of the disease protein.

These starkly visible differences between healthy and diseased cells allowed the neuroscientists to create a highly robust assay for screening compounds that help keep the cells healthy.

Moreover, the assay was an excellent candidate for volume screening of drug candidates. The scientists could image the live cells by a robotic microscope, and then analyze them with advanced image-analysis software to find these changes in structure in an automated and unbiased way. “It takes a lot of the labor out of the work, so we could upscale our platform to screening tens or hundreds of thousands of compounds,” Sepp says.

In a first round of volume screening of chemical compounds, the researchers found four compounds that turned the diseased cells back into normal cells, Sepp says. The compounds also worked well in a fruit-fly model of the disease.

“We wanted to screen many more compounds because we saw how robust the assay was, but the facility we were using was more suited to running smaller drug libraries,” she says. “So that really got us thinking about starting a company.”

Platform Payback
Sepp and Schulte originally envisioned a startup that would offer the screening platform itself. But biotech and business experts in the MIT Enterprise Forum suggested that the platform’s benefits could be demonstrated best by a drug that succeeded in clinical trials.

Following this path, Oxalys is a “virtual company” with the two founders as the only full-time employees. “We will keep it small for the next few years, because as you commercialize a drug it goes very quickly through various stages of drug development that require diverse expertise,” Sepp explains. “The most efficient model is outsourcing, where you bring in consultants and contract research organizations who specialize in these various components as you move along.”

As the startup broadens its funding, Oxalys also looks for backers with expertise. The most suitable venture or pharmaceutical partners will have teams that are experienced in pharmaceutical development, ideally in neurology and more ideally in neurodegeneration, she says.

Moving ahead with pre-clinical development, last year Oxalys achieved the key step of the FDA approval for orphan drug designation for its lead Huntington’s disease therapeutic.

The orphan drug designation brought access to research grants, tax credits for clinical trials and other incentives aimed to smooth the path into clinical studies. Given the rigor of the FDA process, the approval “opened a tremendous number of doors for us,” says Sepp, especially in gaining the attention of leading neurologists around the world who specialize in Huntington’s disease.

Among the neurodegenerative diseases potentially attacked by Oxalys drug candidates, Huntington’s made a logical first target for trials because it is a dominant genetic disorder, driven by mutations of a single gene.

“If one parent has it, then their children have a 50/50 chance of inheriting the disease gene, and if you have the disease gene then you have a 100% chance of developing the disorder,” Sepp says. “We know exactly who will develop the disease, and usually the severity of the disease is similar in these families.” That means that when researchers enlist patients for a clinical trial on Huntington’s, researchers can choose those who will give clearest results on a drug’s effectiveness.

In contrast, scientists can’t predict who will develop Alzheimer’s and how their disease will unfold. “That means you need huge patient populations over a long period of time, which is very cost-intensive,” she notes.

Oxalys is actively developing drug candidates for not only Huntington’s disease but Parkinson’s disease, and hopes to broaden its focus to Alzheimer’s disease as well over time.

The three diseases all are tied to aging, when cell mechanisms such as the ability to control protein quality start to decline, Sepp points out. “If you can find compounds to improve protein quality control, those compounds are likely to be broadly neuroprotective so they will be applicable to all of these major neurodegenerative disorders,” she says. “Treating all of these diseases is our ultimate goal.”



MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.
StartupExchange
October 26, 2015

Creating the Next Line of E-mail Security

Astra IDentity offers holistic solutions for individuals and enterprises to defend against targeted e-mail attacks.
Technology continually updates. Connections get faster. Phones become more powerful. And, unavoidably, computer attacks get smarter. What was once spam as the point of entry has moved to phishing, with hackers elevating their efforts to personalize fake messages.


Gagan Prakash
Co-Founder & CEO

Astra IDentity


Neutralizing this methodology led Gagan Prakash to start Astra IDentity after graduating from the Sloan Fellows Program in 2011. His software compares new email to old ones, developing behavioral fingerprints for each sender to flag suspicious messages. More than providing training programs, his contextual approach solves not only phishing but also the newer problem of spear-phishing, protecting companies’ information, productivity, and profits.

Pointing out the Pretender
Prakash is a software programmer, and, in 2002, he co-founded an IT company in which the flagship product hosted Microsoft Exchange e-mail as a service. After selling it, and, in hopes of finding his next venture, he enrolled in the one-year MIT Sloan Fellows Program in 2010 and got exposed to novel approaches to business and technology problems. One issue was the changing nature of e-mail attacks. Over the previous 10 years, spam had dropped from 90 percent to 70 percent of all e-mail, while the number of large companies attacked with targeted phishing had jumped from 15 to over 50 percent, he says.

With this new data and his previous experience, Prakash found his next venture and co-founded Astra IDentity with his friend and software architect Shyama Gavulla in 2013. Working with their team of engineers, they developed and patented Impostor Detection. The technology determines if a sender’s new e-mails exhibit the same behavior of past ones, looking at hundreds of message characteristics, such as the sender’s email client, location and whitespace style, “thereby forming a dynamic behavioral whitelist rather than the commonly used blacklisting method,” says Prakash, adding that spam filters only screen bad emails sent to hundreds or thousands of people.

By being lower-volume, phishing is more elusive. It also appears less threatening, while being more dangerous, as it looks like it comes from known entities, such as Bank of America or PayPal. Astra IDentity’s software not only catches phishing, but it also focuses on the growing threat of spear-phishing in which the hacker sends out a personalized message using the name and possibly the email address of someone known to the receiver. Given the greater lack of frequency, spam filters particularly struggle with these attacks, Prakash says.



The standard prevention strategy is employee training. Prakash says that the shortcoming in that approach is people soon forget what they’re told, another session is held the following quarter, and the net result is lost work hours without meaningfully increased protection. More than that, the people who have the most to lose, executives, are the most training-averse, partially due to their busy schedules. Prakash says that the Impostor Detection technology is more pro-active, snagging a questionable email before it enters an inbox and either quarantining it, moving it into a junk folder or flagging it for human review. “In essence, we verify if a sender is who they say they are,” he says.

The Downside of Generosity
While anyone with email is a potential customer, Prakash says that he focuses on the commercial market. Consumers are reticent to pay for software, and size plays a factor — attacks happens on those who have large amounts of data. His clients are typically looking to protect trade secrets and sensitive customer and employee information. Along with blanket installations, Prakash says his technology is applied in pockets, such as in human resources, information technology, finance, and the executive department, since these areas have more access to information and are the targets for spear-phishing attacks.

There are a myriad of reasons for breaches. One growing factor is the use of social media. Employees are sharing work-related information across platforms, providing more details for hackers to use in crafting emails, and often doing it on devices that have no installed protection, Prakash says. Not only has it contributed to a rise in spear-phishing, but, in a sector that constantly looks for vulnerability, the over-sharing has also created a new strain of attack, this one on cell phones. The name? Smishing.

While employee training sessions don’t provide automatic filtering, Prakash says that they do teach mindfulness. October is National Cyber Security Awareness Month, and, as part of the effort, Astra IDentity has on its website a 10-tip guide on not getting spear-phished and a simulation called “Can You Spot the Impostor?” With the game, the company generates personalized emails; the visitor tries to determine if the scenario is legitimate. It sounds simple, but Prakash says only 1 out 7 people score perfectly. “I think there’s a little bit of overconfidence in many of the users,” he says. Among the easy-to-miss elements: a different email address for a known sender, a slight misspelling in the sender’s name, and a different address for the link in the body of the email.

But there’s another reason as well. Hackers go to great lengths to achieve authenticity. In one recent instance at a public company, Prakash says that a hacker pretended to be a senior executive and sent targeted messages to the finance department, requesting a $46.7 million transfer. It was approved and only a small fraction of the money is known to have been recovered.

The Campus Effect
While Prakash’s first experience in studying at MIT was at Sloan, the effects were pretty immediate, he says. He ended up taking classes outside the business school and spent time talking to MIT engineers at the Computer Science and Artificial Intelligence Laboratory, along with going to events in the neighboring Kendall Square area. It was in this environment that he was exposed to advances in machine learning and natural language processing. Astra IDentity was formed, he says, by combining the knowledge of these advances with his experience in business e-mail, software as a service, and big data technologies.

Prakash has lived in the Boston suburbs since 1998, but since graduating he’s taken more advantage of endemic MIT talent and resources. “The education is data-centric. It’s very process oriented,” says Prakash, adding that quality is a key benefit where technology information is regularly being produced. It’s especially useful with email security in which new threats quickly outdate established solutions. “People at MIT are able to focus on the details while looking at the big picture,” he says. “We ourselves keep coming up with tactics to prevent the next attacks.”



MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.
StartupExchange
September 28, 2015

Taking Stress to a New Level

MIT spinoff Neumitra brings an innovative approach to quantifying, understanding, and managing stress.
Robert Goldberg never spent much time thinking about fashion. All that changed six years ago, however, when as a visiting neuroscientist at John Gabrieli’s lab in MIT’s Department of Brain and Cognitive Science, he co-founded a Boston-based company called Neumitra. The company develops wearable health sensors and analytics software that provide biofeedback on stress, a condition that Goldberg says costs us $150 billion in lost productivity and more than $190 billion in healthcare costs every year.


Robert Goldberg
Co-Founder & CEO

Neumitra


“I’ve probably looked at over 100,000 wrists over the last few years,” says Goldberg, Neumitra’s CEO. “I look at wrists everywhere I go.”

The experience reinforced a realization that helped define the trajectory of the company. “People don’t all like to wear the same thing,” says Goldberg. He realized that in order to meet Neumitra’s goal of compiling a large data sample for studying stress, and helping as many people as possible gain control over the condition, the technology needed to be integrated into as many wearables as possible. It was unlikely that millions of people would choose to wear a new health bracelet from a small startup.

The cofounders set out to collaborate with a wide variety of wearables manufacturers in order to build their analytics software into health-monitoring devices ranging from smartwatches to fitness bracelets to jewelry. The strategy – and the Neumitra technology – have been a hit with investors. Early investments came from Rock Health, Yahoo, General Magic, Peter Thiel’s Breakout Labs, and more recently, NetScientific.

Startup by Collision
Goldberg met his future cofounders in 2009 at an MIT course called Neurotechnology Ventures. “The class was the kind you could really only find then at MIT,” says Goldberg. “Neumitra was made possible by the collisions between different disciplines that MIT enables.”

Anand Yadav is a biotechnologist, and Safiyy Momen is an algorithms engineer who had been working on ballistic missiles at MIT’s Lincoln Labs. With Goldberg, the trio decided the time was right to develop low-cost devices to measure and analyze stress. Then and now, the key neuroscience research tool was neuroimaging, but the technology still has its limits. “It’s amazing what neuroimaging can show you, but it only gives you about an hour of information at the cost of thousands of dollars,” says Goldberg.

At the time, healthcare wearables were limited and very expensive. Yet Goldberg realized that with increasing miniaturization and dropping sensor prices, the consumer wearables market would soon take off. “Some people told us we had to choose between being a medical or consumer device company. We decided we could be both, and we were proven right. These two worlds are quickly converging.”



The company began developing stress detection algorithms and designed a bracelet prototype called Bandu that primarily depended on skin conductance sensors, also known as galvanic skin response (GSR) or electrodermal activity (EDA). The project led to the realization that “building good hardware is very difficult,” says Goldberg.

The Bandu was scrapped for an improved, Bluetooth enabled Neuma BioWatch, which added ambient temperature and 3-axis motion detection to EDA readings. By the time the Neuma came out, Neumitra had already shifted to the new strategy of focusing on analytics instead of hardware. The company is now developing a series of biomodules using an array of sensors including EDR and heart rate that can be easily integrated into commercial wearables.

“We’re working with device makers to embed our biomodules in devices they’re already building,” says Goldberg. “What we offer at Neumitra is our algorithms, which depend heavily on data fusion, so our hardware partners get better quality data.”

Getting in Touch with Stress
Stress, which reflects the sympathetic “fight-or-flight” part of the autonomic nervous system, “is a basic fact of biology that keeps us alive, but it’s been a problem for us throughout history,” says Goldberg. “Stress affects heart function, breathing, digestion, even fertility. From animal models and neural imaging, we know that stress limits the ability to form new memories, to focus, and to make creative decisions. Under chronic stress, we tend to be less happy and more volatile.”

There are times when stress is advantageous, whether you’re being chased by a lion or cranking on a deadline, but there’s a cumulative effect that takes a toll. “Stress is necessary at times, but the problem really comes from acute and chronic stress. If you confront tough deadlines or social encounters every day, it can whittle away at your best capabilities.”

In addition to selling its biomodules and analytics, Neumitra has a consulting business in which it works with large organizations to help them monitor and analyze the stress of their workers. “We can help companies understand how stress affects different parts of the organization or even one type of role,” says Goldberg.

The conclusions from such research, however, can conflict with business as usual. “One of the biggest challenges we face is the notion that stress is necessary, that you’re not really working hard if you’re not experiencing stress,” he adds. “Yet, organizations are coming to understand that stress affects productivity, long-term healthcare costs, and morale.”

Stress affects most professions and both the rich or poor. Yet, knowledge workers are at a particular disadvantage from chronic stress because they’re expected to use the best capabilities of their brains day in and day out, says Goldberg. Neumitra is working with a software company, for example, to research how stress affects its programmers. “Programming requires a balance between short and long-term memory combined with problem solving skills. When programmers are stressed they are more likely to submit bugs, which are expensive to fix. So this is a cost driver.”

One of Neumitra’s key innovations is its integration of contextual data with the fusion of data from multiple sensors. The Neumitra service provides a smartphone app that links sensor data with contextual information like calendar data and GPS location. The algorithms are also self-learning – if your Neumitra bracelet vibrates to alert you about a high stress level while you’re exercising, you can click on the equivalent of a thumbs down button to instruct Neumitra to ignore a similar signature in the future.

“Stress detection isn’t reducible to one sensor or algorithm, which on their own aren’t sufficient to provide useful feedback,” says Goldberg. “Luckily we all have these computers in our pockets called smartphones. When you combine smartphone data with physiological data, you can learn a tremendous amount about how people, places, and events affect us.”

People are less aware of what stresses them out than they think, says Goldberg. “As a neuroscientist, I thought I understood stress, but I had no clue. I was amazed to find that social situations stress me out. I’m most relaxed when I’m at home quietly working with large data sets.”

While we all have our own particular stress triggers, the data collected from Neumitra users suggests there are certain fairly universal factors, such as participating in a tense meeting or facing a tough deadline. There were some surprises, however, such as the high levels of stress created by eating lunch at your desk while you work.

Just getting to work in the first place can crank you up to a stress level that is hard to dissipate in a workday where meetings and deadlines add to a snowball effect. “The morning commute is one of the most stressful times of the day,” says Goldberg. “People are literally fighting to get into the office on time. When you get there, you’re so overwhelmed by stress, you’re not able to do your best job.”

Neumitra does not presume to tell you whether you should avoid particular stress triggers. “We don’t believe it’s up to us to differentiate between good versus bad stress,” says Goldberg. “Instead, we give the information to the individual and let them decide.”

Goldberg offers the example of people who are stressed out by going on dates. By avoiding such situations, they may lower their stress, but miss out on future happiness. He goes on to note that as the CEO of a startup, he has no choice but to spend much of his day in potentially stressful social situations.

“Sometimes you can’t do much with the feedback, but you can still come back later and address it,” says Goldberg. The mobile app provides users with visualizations and stress management tools such as playing music or games, or suggesting activities like taking a walk. “There are many ways to manage stress, so we try to help users find out what works best for them and when. Sometimes you have time to take a nap, or do meditation or yoga, but other times you only have enough time to play a song.”

The Neumitra wearables can be worn day and night, thereby revealing connections between stress and sleep that “were a real surprise to me,” according to Goldberg. “We’re beginning to see how stress affects the way you sleep, and how the lack of sleep affects the stress you feel the following day. Stress and sleep are two sides of the same physiological coin: the sympathetic vs. the parasympathetic nervous system. They are constantly at tension and not always in ways that we understand.”

Boston Stress Study to Quantify City’s Stress
As the number of Neumitra users grows, the company is learning a lot about the differences in stress among organizations, professions, genders, socioeconomic classes, and even countries. “It surprised me that stress rates are not only higher in cities than rural areas, but higher in some countries compared to others,” says Goldberg. “Ultimately, that’s how we’re going to solve stress, by looking at data across all humanity.”

As a next step, Neumitra is launching a Boston Stress Study in late 2015, with first results expected in early 2016. The company has already signed up over 1,000 Bostonians to wear Neumitra bracelets, with plans to expand that. The participants include a diverse range of roles, from students to police officers to CEOs. Partners include hospitals, companies, and other organizations, such as stressed out universities like MIT and Harvard.

“We’re working with a wide range of partners to help make it a much bigger study,” says Goldberg. “Our model is the Framingham Heart Study, which started in 1948 with 5,209 people. It’s now on its third generation, and has taught us pretty much everything we know about the heart.”

The Boston Stress Study will examine how stress affects the city of Boston at large, as well as specific organizations, both in real time and long term. “We will develop a map of how and when we experience stress by commute type, profession, location, gender,” says Goldberg. “By working with us on the study you’ll not only learn how stress is affecting your organization, but how stress affects us all. For the first time, we will begin to understand how our daily lives are driving ourselves just a little bit crazy.”



MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.
StartupExchange
July 13, 2015

Computing at Full Capacity

Jisto helps companies optimize utilization across all available computing resources with real-time deployment, monitoring, and analytics.
According to a 2014 study from NRDC and Anthesis, in 2013 U.S. data centers burned 91 billion kilowatt-hours of electricity, enough to power every household in New York City twice over. That figure is expected to rise to 140 billion by 2020. While improved energy efficiency practices could go a long way toward lowering this figure, the problem is greatly exacerbated by the underutilization of servers, including an estimated 30 percent of servers that are still plugged in, but are no longer performing any services, says the study.


Aleksandr Biberman
Co-founder & CEO

Jisto Inc.


In another 2014 study, tech research firm Gartner, Inc. found that data center systems collectively represent a $143 billion market. With enterprise software adding another $320 billion and IT services another $963 billion, the overall IT industry represents a whopping $3.8 trillion market.

Companies are increasingly seeking new ways to cut costs and extract the largest possible value from their IT infrastructure. Strategies include placing data centers in cooler climates, switching to more affordable open source software, and virtualizing resources to increase utilization. These solutions just scratch the surface, however.

An MIT-connected startup called Jisto offers businesses an interesting new tool for cutting data center and cloud costs while improving resource utilization. Jisto manages existing enterprise applications by automatically wrapping them in Jisto-managed Docker containers, and intelligently deploying them across all available resources using automated real-time deployment, monitoring, and analytics algorithms. As the resource utilization profile changes for each server or different parts of the network and storage, Jisto elastically scales its utilization in real time to compensate.

“We’re helping organizations get higher utilization of their data center and cloud resources without worrying about resource contention,” says Jisto CEO and Co-Founder Aleksandr (Sasha) Biberman. So far, the response has been promising. Jisto was a Silver Winner in the 2014 MassChallenge, and early customers include data-intensive companies like banks, pharmas, biotechs, and research institutions.



“There’s pressure on IT departments from two sides: how can they more efficiently reduce data center expenditures, and how can they improve productivity by giving people better access to resources,” says Biberman. “In some cases, Jisto can double the productivity with the same resources just by making better use of idle capacity.”

Biberman praises the MIT Industrial Liaison Program and Venture Mentoring Service for hosting networking events and providing connections. “The ILP gave us connections to companies that we would have never otherwise have connected to all around the world,” he says. “It turned us into a global company.”

Putting Idle Servers Back to Work
The idea for Jisto came to Biberman while he was a postdoc in electrical engineering at MIT Research Lab of Electronics (RLE), studying silicon photonic communications. While researching how optical technology could improve data center performance and efficiency, he discovered an even larger problem: underutilization of server resources.

“Even with virtualization, companies use only 20 to 50 percent of in-house server capacity,” says Biberman. “Collectively, companies are wasting more than $100 billion annually on unused cycles. The public cloud is even worse, where utilization runs at 10 to 40 percent.”

In addition to the problem of sheer waste, Biberman also discovered that workload resources are often poorly managed. Even when more than a half of a company’s resources are sitting idle, workers often complain they can’t get enough access to servers when they need them.

Around the time of Biberman’s realization, he and his long-time friend Andrey Turovsky, a Cornell-educated tech entrepreneur, and now Jisto CTO and Co-Founder, had been brainstorming some startup ideas. They had just developed a lightweight platform to automatically deploy and manage applications using virtual containers, and they decided to apply it to the utilization and workload management problem.

Underutilization of resources is less a technical issue, than a “corporate risk aversion strategy,” says Biberman. Companies tend to err on the side of caution when deploying resources and typically acquire many more servers than they need.

“We started seeing some crazy numbers in data center and cloud provisioning,” said Biberman. “Typically, companies provision for twice as much as they need. One company looks at last year’s peak loads, and overprovisions above that by a factor of four for the next year. Companies always plan for a worst-case scenario spike. Nobody wants to be the person who hasn’t provisioned enough resources, so critical applications can’t run. Nobody gets fired for overprovisioning.”

Despite overprovisioning, users in most of the same organizations complain about lack of access to computing resources, says Biberman. “When you ask companies if they have enough resources to run applications, they typically say they want more even though their resources are sitting there going to waste.”

This paradox emerges from the common practice of splitting access into different resource groups, which have different levels of access to various cluster nodes. “It’s tough to fit your work into your slice of the pie,” says Biberman. “Say my resource group has access to five servers, and it’s agreed that I use them on Monday, and someone else takes Tuesday, and so on. But if I can’t get to my project on Monday, those servers are sitting completely idle, and I may have to wait a week. Maybe the person using it on Tuesday only needs one of the five servers, so four will sit idle, and maybe the guy using it the next day realizes he really needs 10 or 20 servers, not just the five he’s limited to.”

Jisto breaks down the artificial static walls created with ownership profiles and replaces them with a more dynamic environment, says Biberman. “You can still have priority during your server time, but if you don’t use it, someone else can. That means people can sometimes get access to more servers than were allotted. If there’s a mission-critical application that generates a spike we can’t predict, we have an elastic method to quickly back off and give it priority.”

Financial services companies are using Jisto to free up compute cycles for Monte Carlo simulations that could benefit from many more servers and nodes. Pharma and life science companies, meanwhile, use a similar strategy to do faster DNA sequencing. “The more nodes you have, the more accurately you can run a simulation,” says Biberman. “That’s a huge advantage.”

Docker Containers for the Enterprise
Jisto is not the only cloud-computing platform that claims to improve resource utilization and reduce costs. The problem with most, however, is that “if you have a really quick spike in workload, there’s not enough time to make intelligent decisions about what to do,” says Biberman. “With Jisto, an automatic real-time decision-making process kicks in, enabling true elasticity across the entire data center with granularity as fine as a single core of a CPU.”

Jisto not only monitors CPU usage but other parameters such as memory, network bandwidth, and storage. “If there’s an important memory transfer happening that requires a lot of bandwidth, Jisto backs off, even if there’s plenty of CPU power available,” says Biberman. “Jisto can make intelligent decisions about where to send jobs based on all these dynamic factors. As soon as something changes, Jisto decides whether to stop the workload, pause it, or reduce resources. Do you transfer it to another server? Do you add redundancy to reduce the latency tail? People don’t have to make and implement those decisions.”

The platform also integrates rigorous security provisions, says Biberman. IT directors are understandably cautious about bringing third-party software into their complex data center ecosystems, which are often protected by firewall and regulation settings. Jisto, however, can quickly prove with a beta test how the software can spin its magic without interfering with mission-critical resources, he adds.

Jisto’s unobtrusiveness is largely due to its use of Docker containers. “Docker has nice APIs and makes the process much easier, both for us as developers and for Jisto customers,” says Biberman. “Docker is very portable—if you can run it on Linux, you can run it on Docker—and it doesn’t care if you’re running it on a local data center, a private cloud, or on Amazon. With containers, we don’t need to do something complicated like run a VM inside another VM. Docker gives us a lightweight way to let people use the environment that’s already set up.”

Based in Cambridge, Massachusetts, Jisto was the first, and remains one of few, Docker-based startups in this region.

Moving Up to the Cloud
Companies are increasingly saving on data center costs by using public cloud resources in a hybrid strategy during peak demand. Jisto can help bridge the gap with better efficiency and flexibility, says Biberman. “If you’re a bank, you might have too many regulations on your data to use the public cloud, but most companies can gain efficiencies with public clouds while still keeping their private cloud for confidential, regulated, or mission-critical tasks.”

Jisto operates essentially the same whether it’s running on-premises, or in a private, public, or hybrid cloud. Companies that exceed the peak level of their private data center can now “burst out” onto the public cloud and take advantage of the elastic nature of services like Amazon, says Biberman. “Some companies provision hundreds of thousands of nodes on Amazon,” he adds. The problem is that Amazon charges by the hour. “If a company only needs five minutes of processing, as many as 100,000 nodes would sit idle for 55 minutes.”

Jisto has recently begun to talk to companies that do cloud infrastructure as a service, explaining how Jisto can reprovision wasted resources and let someone else use them. It’s only a matter of time before competitive pressures lead a cloud provider to use something like Jisto, says Biberman.



MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.
StartupExchange
June 8, 2015

Text, Context, and Insight

Luminoso brings adaptable contextual awareness and common sense to automated text analytics for insight into digital conversations of all kinds.
When navigating our proliferating digital interfaces, one might be forgiven the suspicion that humans are out of the loop in a society run by computers. Yet consider for a moment how paranoid computers must feel.


Catherine Havasi
Co-founder & CEO

Luminoso



Human communications are filled with nuance, metaphor, jargon, double entendre, context, and murky oceans of assumed knowledge. When interpreting speech or text, especially the casual chitchat of social networking, computers are often clueless.

A text analytics software company called Luminoso aims to keep computers in the human loop. “We help computers understand people better and help companies understand their customers better,” says CEO and co-founder Catherine Havasi. “Companies are trying to understand the text their customers generate and how they talk about things like flavors, senses, colors, and textures. They want to do this in a quantitative way so they can make data-driven decisions.”

When she’s not working at Luminoso, Havasi helps improve the “ConceptNet” open data model project behind the software at the MIT Media Lab in her role as research scientist. The ConceptNet database currently stores some 17 million concepts in all major languages in a way that makes sense to computers.

Luminoso’s SaaS-based Analytics Platform builds on ConceptNet with analytics tools and visualizations that help build actionable insights from unstructured text data. Primarily sold to marketing departments, the software analyzes text derived from Twitter feeds, email, survey forms, tech support logs, and other inputs. Luminoso recently released a solution called Compass that analyzes text in real time.

From Search to Text Analytics
ConceptNet wasn’t originally focused on marketing but rather on building a better search engine. In the late ‘90s, Havasi was working at the Media Lab with Marvin Minsky’s Society of Mind group when it set out to help computers interpret human search input.



“Internet search had just gotten started, and people tended to type in statements like ‘My cat is sick’ and would get nothing helpful in return,” says Havasi. To help bridge the gap, the group launched the Open Mind Common Sense Project (OMCS), whose mission was “to collect all the things people know but computers don’t,” says Havasi. This common sense or “world knowledge” includes relationships between physical objects or attributes, as well as human motivations.

Over the years, the search problem was largely solved, primarily by people learning how to pose questions the way a computer might. Yet, OMCS realized that teaching computers assumed knowledge had much broader applicability, especially in text analytics.

Common sense statements had never been collected in a comprehensive way, and the task seemed daunting. To speed the process, OMCS developed a technique, later known as crowdsourcing, which they referred to as “harnessing the power of bored people on the Internet,” says Havasi. “We put up a public web page with an input box and a prompt that said ‘Teach the computer.’ Each morning, we checked the statements, and decided which ones were true. Later, we developed automated ways to check the knowledge.”

As people found better things to do on the Internet, the OMCS motivated participants by integrating the inputs into games. “People would play Verbosity and teach the system without thinking about it,” says Havasi.

The researchers enhanced the resulting ConceptNet database by integrating other world knowledge. For example, they taught the computer to automatically scan Wikipedia and import common sense relationships.

ConceptNet Gets Multi-lingual and Multi-cultural
Much of the current ConceptNet research involves encapsulating how common sense is perceived and expressed in different cultures and languages. “How we think of a cat can be very different depending on your culture,” says Havasi. “The same goes for the way we describe what we want from a hotel experience or even the concepts of flavor or the expected hour for dinner. You need to add cultural and linguistic nuance.”

Even when the task is limited to a single language, computers are challenged by passionate, creative, or playful text. “We use all these metaphors and try to say things in new and interesting ways, especially online, because we want people to listen to us,” says Havasi. “Our world knowledge helps us understand this language, and it can do the same for computers.”

One company asked Luminoso to help them decipher a customer comment claiming a product smelled “musty,” relates Havasi. “They needed to understand whether this was something isolated or systematic.” A typical text analytics system would stop after searching for the word “musty” or other synonyms. However, ConceptNet and Luminoso might extend that to notice a post saying “the product smells like an old house,” she adds.

Launching Luminoso
The idea for Luminoso emerged when Havasi was directing the MIT Media Lab’s Digital Intuition group, working with member companies to help apply ConceptNet to text analytics. She quickly realized ConceptNet alone could not serve typical business needs.

“ConceptNet showed promising results, but companies wanted to use it in a more data-driven way,” she says. These insights resulted in the launch of Luminoso in 2010.

Typical text analytics solutions fail to meet the needs of marketing departments, says Havasi. “Statistical software that does things like look for how often words appear together requires a lot of data, and hand-coded ontology, or ruleset, systems lack adaptability,” says Havasi. “As the world changes along with the way people talk, the ontology can’t keep up, especially since it usually requires manual updates. You’re taking the person out of one part of the process and putting him into another part, which really isn’t solving the problem.”

By contrast, Luminoso’s Analytics Platform integrates self-learning algorithms that reduce the need for human updates. The software analyzes text ranging from survey open-ends to social media logs and builds actionable insights that can be applied across the product lifecycle.

“We can look at things that are hard to pin down, like intent to purchase or openness to new technology, or what improvements might help people advocate for a product,” says Havasi. “Our software can help figure out whether or not a brand can authentically build a certain kind of product or answer questions like what kind of SKUs drive somebody into a retail store.”

The Analytics Platform examines how a word is used, and then makes relations or analogies with the way other words are used. The software also finds contextual clues in customer metadata, geographical region, or the time a tweet was sent. The software can identify the number of stars a customer clicked on for a hotel review, and stir that into the story, or look at Twitter feeds to guess whether someone is a Republican or Democrat, says Havasi.

Marketing departments have invested heavily in analyzing social media, yet they often miss out on other text sources, says Havasi. “There’s incredibly rich information about your product online that is not in social media,” she says. “For example, there are forms devoted to shaving or cars, or how to optimize airline travel. One of the Analytics Platform’s strengths is that it doesn’t matter where the text comes from.”

Compass: Real-time Text Analysis
In early 2015, Luminoso released Compass, which analyzes streams of text data in real time. “Compass lets you understand trends as they evolve, giving you early warning,” says Havasi.

In 2014, Sony used a Compass prototype when sponsoring the World Cup. Sony’s digital agency Isobar built a second screen experience called One Stadium Live that let fans watch World Cup updates on a tablet, and then comment via social media. It was the largest social media event in history.

The huge volume of “big text” generated by One Stadium Live was a challenge for Luminoso, but it was not so much the volume as the variability, says Havasi. “If someone scored a goal, everybody tweeted about it, which could overwhelm the experience. We needed to find who was contributing uniquely and interestingly, and determine which topics people cared about.”

Compass also needed to respond dynamically to new developments. When Luis Suárez of Uruguay bit Giorgio Chielleni of Italy, the topic of biting “suddenly became relevant to soccer, which was something we couldn’t anticipate,” says Havasi. “Compass had to react quickly.” For the most part, Compass was able to do that without human intervention, thereby greatly reducing reaction time.

Plugging In, Reaching Out
Moving forward, Luminoso will continue to enhance its algorithms to help computers get the gist of ambiguous human communications. Yet, much of the focus is now on customer integration, including back-end IT plumbing and graphical visualizations.

Luminoso recently released APIs for the Analytics Platform that let users classify and tag text data without using the software’s GUI. In this way, customers can integrate the analytics into their existing software.

The software is currently a cloud-only platform, but Luminoso will soon release an on-premises version that will “go behind firewalls or handle highly secure data,” says Havasi. “This should open up new markets like financial services, pharmaceuticals, and law enforcement.”

A growing focus for Luminoso is to effectively communicate the sometimes subtle insights from the Analytics Platform or Compass. Like Luminoso’s core analytics, this process involves bridging the communications gap between computers and people.

“Visualizing the conclusions is almost as hard as coming up with the conclusions themselves,” says Havasi. “We’re always searching for different ways of visualizing our data, for example using different types of word clouds and heat maps, so that people can use it to make actionable decisions. We need to convey the conclusions to people all over the company, not only to analysts.”



MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.