Autonomous systems such as robots, drones, and autonomous vehicles are increasingly becoming a common element of human society. However, how to make autonomous systems more intelligent, more efficient, safer, and better aligned with human values remains to be a big challenge. The MIT Schwarzman College of Computing invests in new artificial intelligence, data science, and computer science research to advance the state of the art in autonomous systems. The 2023 MIT AI and Autonomy Conference will showcase the most recent developments in areas including networked robotics, human level perception, reinforcement learning, and computational cognitive science. Presented by the MIT Schwarzman College of Computing and the MIT Industrial Liaison Program, you will hear insights from MIT faculty, MIT Startup Exchange entrepreneurs, and industry executives from presentations, lightning talks, exhibitions, and a panel discussion.
John Roberts has been Executive Director of MIT Corporate Relations (Interim) since February 2022. He obtained his Ph.D. in organic chemistry at MIT and returned to the university after a 20-year career in the pharmaceutical industry, joining the MIT Industrial Liaison Program (ILP) in 2013. Prior to his return, John worked at small, medium, and large companies, holding positions that allowed him to exploit his passions in synthetic chemistry, project leadership, and alliance management while growing his responsibilities for managing others, ultimately as a department head. As a program director at MIT, John built a portfolio of ILP member companies, mostly in the pharmaceutical industry and headquartered in Japan, connecting them to engagement opportunities in the MIT community. Soon after returning to MIT, John began to lead a group of program directors with a combined portfolio of 60-80 global companies. In his current role, John oversees MIT Corporate Relations which houses ILP and MIT Startup Exchange.
Hong Fan is a Program Director at the Office of Corporate Relations at MIT. She joined OCR in August 2016, brought with her 20+ years of international work experience across semiconductor, consumer electronics, telecom, and higher education.
Prior to joining OCR, Hong spent 12 years in the semiconductor industry with executive functions in strategic marketing, business development, corporate strategy, product management, and product marketing at Analog Devices and MediaTek. During those years, Hong played instrumental roles in identifying emerging business opportunities related to wireless communication networks, smartphones, wearable devices, Internet of Things (IoT), and medical devices and applications. She led cross-functional teams in defining and driving product and market strategy for businesses with annual revenue ranging from $30 million to $100 million.
Prior to joining the semiconductor industry, Hong spent 6 years in the telecommunications and electronics industry, leading engineering teams at companies such as Lucent Technologies and Watkins-Johnson Company for the development of digital signal processing, wireless communications, and micro-controller software.
Before coming to US, Hong was a strategic research staff at the President Office of Shanghai Jiao Tong University, one of the oldest universities in China. She was the first woman to hold this highly selective position.
Hong has a B.S in Electronic Engineering from Shanghai Jiao Tong University, an M.S. in Electrical Engineering from University of Maryland at College Park, and an MBA from Sloan School of Management at MIT. She received numerous academic honors and awards including the McKinsey & Co. Scholarship, the NSF Graduate Research Fellowship, and the Shanghai Outstanding College Graduate Award.
Daniel Huttenlocher is the inaugural dean of the MIT Stephen A. Schwarzman College of Computing. He began his academic career at Cornell University in 1988, where he was a member of the computer science faculty. In 1998, he chaired the task force that led to the creation of Cornell’s interdisciplinary Faculty of Computing and Information Science, later serving as its dean starting in 2009. In 2012, he became the founding dean of the new Cornell Tech campus in New York City.
Huttenlocher has extensive industry experience, having served as a scientist and lab director at Xerox’s Palo Alto Research Center for 12 years before leaving to help establish a financial technology startup, Intelligent Markets, in 2000.
Huttenlocher’s research and scholarship in computer science is broad and interdisciplinary, spanning algorithms, social media, and computer vision. He has earned the Longuet-Higgins Award for Fundamental Advances in Computer Vision (2010), and various fellowships and awards from the National Science Foundation, the Association for Computing Machinery, IEEE, and Phi Beta Kappa.
He is a member of the boards of directors of Amazon and Corning, and of the John D. and Catherine T. MacArthur Foundation, where he has served as chair since 2018.
Huttenlocher earned a bachelor’s degree from the University of Michigan in 1980, double-majoring in computer and communication sciences and experimental psychology. An MIT alumnus, he earned an SM in electrical engineering and computer science in 1984 and a PhD in computer science in 1988.
Generative AI raises unparalleled questions. It may be tempting to see it as another in a long line of information and communication technologies, which have resulted in the profusion of human expression. However, generative AI operates in a very different manner than prior technologies, distilling human expression in order to synthesize human-like expression. The outcome of such a process, while not yet well understood, is qualitatively different from that of the human mind, creating gaps with human understanding. The essential challenges and opportunities of generative AI are thus cognitive at least as much as technological. Navigating this transformation successfully will require new concepts of human thought, not only new modes of human interaction with machines.
Wojciech Matusik is a professor in MIT's Department of Electrical Engineering and Computer Science, and leads the Computational Fabrication Group at the Computer Science and Artificial Intelligence Laboratory. His research interests are in computer graphics, computational design and fabrication, computer vision, robotics and human-computer interaction. Before coming to MIT, he worked at Mitsubishi Electric Research Laboratories, Adobe Systems and Disney Research Zurich. He has received a Ruth and Joel Spira Award for Excellence in Teaching, a DARPA Young Faculty Award and a Sloan Foundation fellowship. He has been named one of the world's top 100 young innovators by MIT Technology Review and received a Significant New Researcher Award from ACM Siggraph. He earned a PhD in computer graphics at MIT.
Design is everywhere: high-performance turbines, polymers with
outstanding material properties, unmanned aerial vehicles,
metamaterials, or computer algorithms. However, the best designs are a
product of tremendous work of high-skilled domain experts. I will show
that we are on the verge of a transition where computational methods
start beating humans at design. I will describe a series of questions
that need to be addressed to move the field of computational design
forward: how to represent a design, how to represent design space, how
to find designs with optimal performance, and how to bridge the gap
between simulation and reality.
Russ Tedrake is the Toyota Professor in MIT's Department of Electrical Engineering and Computer Science, and a member of the Computer Science and Artificial Intelligence Laboratory, where he leads the Center for Robotics. He is also vice president of robotics research at the Toyota Research Institute. His research focused on motor control systems in animals and machines that can execute dynamically dexterous tasks and interact with uncertain environments. Current projects include robust and efficient bipedal locomotion on flat terrain, multi-legged locomotion over extreme terrain, flapping-winged flight, and feedback control for fluid dynamics. He has received an NSF Career Award, an MIT Jerome Saltzer Award for undergraduate teaching, a DARPA Young Faculty Award, an MIT Spira teaching award, and a faculty fellowship at Microsoft Research. He earned a BSE from the University of Michigan, Ann Arbor, and a PhD from MIT.
Advances in machine learning and control have enabled a new breed of dexterous robots that can leverage rich perceptual inputs and manipulate objects that would be difficult to model in a simulation. In this talk, I'll describe recent results in feedback control from pixels, intuitive physics, and control for dexterous manipulation. I'll describe some open theoretical challenges and show some fun experimental results.
Vivienne Sze is an Associate Professor in the Electrical Engineering and Computer Science Department at MIT. She works on computing systems that enable energy-efficient machine learning, computer vision, and video compression/processing for a wide range of applications, including autonomous navigation, digital health, and the internet of things. She is widely recognized for her leading work in these areas and has received many awards, including the AFOSR and DARPA Young Faculty Award, the Edgerton Faculty Award, several faculty awards from Google, Facebook, and Qualcomm, the 2018 Symposium on VLSI Circuits Best Student Paper Award, the 2017 CICC Outstanding Invited Paper Award, and the 2016 IEEE Micro Top Picks Award. As a member of the JCT-VC team, she received the Primetime Engineering Emmy Award for the development of the HEVC video compression standard. She is a co-editor of High Efficiency Video Coding (HEVC): Algorithms and Architectures (Springer, 2014) and co-author of Efficient Processing of Deep Neural Networks (Synthesis Lectures on Computer Architecture, Morgan Claypool, 2020). For more information about Prof. Sze’s research, please visit http://sze.mit.edu.
A broad range of next-generation applications will be enabled by low-energy autonomous vehicles including insect-size flapping wing robots that can help with search and rescue, chip-size satellites that can explore nearby stars, and blimps that can stay in the air for years to provide communication services in remote locations. Autonomy capabilities for these vehicles will be unlocked by building their computers from the ground up, and by co-designing the algorithms and hardware for autonomy and navigation. In this talk, I will present various methods, algorithms, and computing hardware that deliver significant improvements in energy consumption and processing speed for tasks such as visual-inertial navigation, depth estimation, motion planning, mutual-information-based exploration, and deep neural networks for robot perception. We will also discuss the importance of efficient computing to reduce carbon emissions for sustainable large-scale deployment of autonomous vehicles. Much of the work presented in this talk was developed in the Low-Energy Autonomy and Navigation (LEAN) interdisciplinary group at MIT (http://lean.mit.edu), which is co-directed by Vivienne Sze and Sertac Karaman.
Ariadna Rodenstein is a Program Manager at MIT Startup Exchange. She joined MIT Corporate Relations as an Events Leader in September 2019 and is responsible for designing and executing startup events, including content development, coaching and hosting, and logistics. Ms. Rodenstein works closely with the Industrial Liaison Program (ILP) in promoting collaboration and partnerships between MIT-connected startups and industry, as well as with other areas around the MIT innovation ecosystem and beyond.
Prior to working for MIT Corporate Relations, she worked for over a decade at Credit Suisse Group in New York and London, in a few different roles in event management and as Director of Client Strategy. Ms. Rodenstein has combined her experience in the private sector with work at non-profits as a Consultant and Development Director at New York Immigration Coalition, Immigrant Defense Project, and Americas Society/Council of the Americas. She also served as an Officer on the Board of Directors of the Riverside Clay Tennis Association in New York for several years. Additionally, she earned her B.A. in Political Science and Communications from New York University, with coursework at the Instituto Tecnológico y de Estudios Superiores de Monterrey in Mexico City, and her M.A. in Sociology from the City University of New York.
Dr. Cyrus Shaoul obtained his BSc in Cognitive Science from MIT in 1993, and his PhD from the University of Alberta in 2012. While at MIT he was an undergraduate researcher at the Media Lab. He co-founded one of the first Internet technology companies in Japan in 1994, called Digital Garage. Cyrus is a co-founder and the CEO of Leela AI. He and his team are laser-focused on building a scalable solution to the problem of building a new, more powerful kind of visual intelligence. Cyrus speaks Japanese, French and Spanish.
Javier is a Co-Founder at Inkbit, a Boston company that develops industrial additive manufacturing systems for a variety of application areas including robotics, fluidics, and aerospace. At Inkbit, he is involved in technology and application development. Previously, he was a research engineer and a graduate student at the MIT Computational Fabrication Group where he developed metrology systems for in-situ inspection and closed-feedback correction of 3D printed parts. The technology developed at MIT served as the technical basis for Inkbit. He holds a BS and MS in mechanical engineering from MIT.
Sebastian is the CEO of ubicept which provides new imaging solutions in challenging environments, such as seeing motion in the dark at unprecedented quality. He has spent his career exploring new kinds of imaging, focusing on signal processing and information extraction. He holds BS/MS/Ph.D. degrees in Electrical Engineering from the Karlsruhe Institute of Technology (KIT), Germany.
Matthew Cherewka serves as Director of Product Marketing and Strategy at Vecna Robotics, where he leads strategic go-to-market, product, and growth initiatives. Prior to this, he led Business Development for the logistics vertical, and helped spin out Vecna Robotics from an R&D lab into a venture-backed independent startup.
Matt is an avid evangelist for Industry 4.0, and has spent over a decade in the robotics and autonomy industries consulting for leading global enterprises on automation strategies and working with startups to commercialize new robotics technologies. He is an alumni of Bucknell University with degrees in mechanical engineering and business management, and brings a diverse set of experiences across product development, applications engineering, enterprise solutions sales, management consulting, and marketing strategy to help bridge the gap between developers and users of automation technology.
Outside of work, Matt enjoys spending time with his family, friends and pets, and plays fiddle with several bands across Greater Boston and New England.
Sampriti Bhattacharyya is the co-founder and CEO of Navier, a bay area startup building the next generation maritime company. She is a roboticist and received her PhD in Mechanical Engineering specializing in Robotics and Control Theory from MIT (2017) along with a minor in business. At MIT, Sampriti worked on underwater drones and mathematical modeling of nonlinear dynamical systems which formed the foundation of the project Hydroswarm. Prior to MIT, Sampriti earned a MS in Aerospace Engineering at The Ohio State University and worked on Intelligent Flight Control Systems at NASA Ames. Sampriti's recent work has been extensively focussed on understanding of technology explosion in the maritime space, market opportunities, and she has given many talks both in academic and industry-related conferences like Forbes Under 30, CERAWeek, TTI Vanguard, Oceans Conference, Techcrunch, and others.
As a masters student, Sampriti worked with Fermilab, a DOE High Energy Physics Lab in Chicago on a new kind of innovative nuclear reactor: Accelerator Driven Subcritical Reactor for producing energy out of radioactive waste. She was also a Research Fellow at Tata Institute of Fundamental Research (TIFR), one of India’s top research organizations. She completed her bachelors in Electrical Engineering in India and worked on several renewable energy projects funded by the Indian Department of Science and Technology. Her interdisciplinary background enables her to address large complex problems in frontier tech like robotics, AI, large data-driven systems, and those at the intersection of policy, security and technology.
She has received many accolades from being recognized as one of the Forbes 30 under 30 (2016), Robohub’s 25 Women in Robotics, MassChallenge Gold Prize Winner among many others. Her invention, the Hydrone has been lauded at the Senate Invention Caucus and is now on display in the Victoria & Albert Design Museum in London as one of the top 100 projects shaping the future. She is an active advocate for women in STEM fields and was also awarded MIT’s most outstanding mentor award, 2016.
Lars Erik's first encounter with startups was at the age of 21 when his first startup, Gobi, raised a round from American investors, and he dropped out of college to continue the pursuit in Palo Alto, California. The startup is currently valued at 20 $million, but Lars Erik sold his shares to pursue more advanced technology. After two internships at Lockheed Martin Solar and Astrophysics Laboratory in Palo Alto and one internship at Palantir in London, he met the co-founders of Aviant while studying at MIT. Since then, his sole focus is with the company, which has grown very quickly and currently operates three commercial contracts for hospitals in Scandinavia.
Hector is the CEO of Rotor, a 25-person startup based in the Boston area building a teleoperations system that makes flying helicopters safer and more economical -- Rotor's helicopters are being used to fight wildfires as soon as later this year. Hector obtained his PhD at the Laboratory for Aviation and the Environment in the MIT AeroAstro department and continues to contribute to research shaping the future of aviation and sustainability.
Ian founded AdaViv while holding a Postdoctoral Associate position at MIT and his background and interests bridge plant science, agriculture, image processing and data science. As an entrepreneur and technologist at heart, Ian is guiding AdaViv's mission to help every indoor farm become highly profitable and sustainable through enhanced Plant-Intelligence and Lean Cultivation tools.
Vinayak Ramesh is the CEO and co-founder of the MIT startup, Ikigai Labs, a platform for building AI Apps. Previously, he co-founded and was CTO of Wellframe, a digital health company focused on AI to help payors better manage their patient populations (acquired by Blackstone). He received his S.B./M. Eng degrees from MIT and is Forbes 30 under 30.
Fadi Micaelian, founder and CEO of Sparkdit, is an entrepreneur with a seasoned background in enterprise software. He served over the past three decades in several executive positions at industry leaders like Oracle, BroadVision, Intellectual Ventures, ViewStar (now OpenText), DataBeam (now IBM) and Element Data. Most prominently, he was the founder of Auguri Corporation where he served as its CEO for 13 years, leading it to a successful acquisition.
Fadi is also a distinguished scholar and an inventor, who has been repeatedly published, and was awarded several patents by the USPTO, in AI/ML, Cloud, Big Data, Analytics, Decision and IoT. Author of early inventions in the field of AI/ML, his patents, considered essential, rated 90-100 percentile are highly cited by Google, Microsoft, eBay, Oracle, EMC, Yahoo, GE, etc.
He is passionate about Technology, Physics and Track and Field. Fadi holds a BS of Physics, a BE of Engineering, and obtained an MS in Engineering from MIT, and an MBA from INSEAD.
Aude Oliva, PhD is the MIT director in the MIT-IBM Watson AI Lab and director of strategic industry engagement in the MIT Schwarzman College of Computing, leading collaborations with industry to translate natural and artificial intelligence research into tools for the wider world. She is also a senior research scientist at the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL), where she heads the Computational Perception and Cognition group. Oliva has received an NSF Career Award in computational neuroscience, a Guggenheim fellowship in computer science and a Vannevar Bush Faculty Fellowship in cognitive neuroscience. She has served as an expert to the NSF Directorate of Computer and Information Science and Engineering on the topic of human and artificial intelligence. She is currently a member of the scientific advisory board for the Allen Institute for Artificial Intelligence. Her research is cross-disciplinary, spanning human perception and cognition, computer vision and cognitive neuroscience, and focuses on research questions at the intersection of all three domains. She earned a MS and PhD in cognitive science from the Institut National Polytechnique de Grenoble, France.
Sertac Karaman is the director of the Laboratory for Information and Decision Systems, and an associate professor of Aeronautics and Astronautics at MIT. His research areas are robotics and control theory, particularly the applications of probability theory, stochastic processes, stochastic geometry, formal methods, and optimization for the design and analysis of high-performance cyber-physical systems. The applications of this research include driverless cars, unmanned aerial vehicles, distributed aerial surveillance systems, air traffic control, and certification and verification of control systems software. Karaman received a PhD in electrical engineering and computer science and an SM in mechanical engineering from MIT and BS degrees in mechanical engineering and in computer engineering from the Istanbul Technical University.
Powered aerial vehicles have been around for more than a century. They have always been operated by a dedicated pilot, who has been on the vehicle for most of the century; remotely-piloted vehicles were introduced relatively recently, and still require, often multiple, vehicles. Today, the field of aerial vehicles is going through one of the most exciting breakthroughs in its hundred-year history. Autonomy-enabled vehicles will transform not only the aerospace industry but also consumer electronics, inspection and services, and more. In this talk, we discuss the future of autonomy-enabled aerial vehicles. We highlight key emerging technologies, including agile vehicles, miniature vehicles, vehicle teams and others.
John Tylko serves as Aurora’s Chief Innovation Officer, responsible for leading technology strategy and commercialization efforts. He oversaw Aurora’s major new program and customer acquisition efforts, enabling Aurora’s rapid growth which led to its acquisition by Boeing in 2017. He also developed the vision for Boeing’s Aerospace and Autonomy Center.
His career has been focused on the development of innovative technologies and products, spanning both the aerospace and the electronics and computer industry. He received a Bachelor of Science in Aeronautics and Astronautics from the Massachusetts Institute of Technology in 1979 and began his engineering career at General Electric where he developed the first composite structural assemblies for aircraft engines. Tylko cofounded General Computer in 1981, which developed innovative electronics and personal computer products. He also cofounded VideoGuide, which developed the first interactive television program guide.
Tylko helped start Aurora Flight Sciences with John Langford in 1989 and was a member of Aurora’s Board of Directors for nearly thirty years from 1989 to 2017. He led Aurora’s Global Hawk program, built its state of-the-art composite structures manufacturing center, and established Aurora’s aerostructures business sector. Tylko founded Aurora’s Research and Development Center in Cambridge in 2005 and was the principle architect of Aurora’s strategy for aerospace autonomy.
He has been a lecturer in MIT’s Department of Aeronautics and Astronautics and is a recipient of MIT’s Founders Award which recognizes innovation and entrepreneurship. Tylko was a Guggenheim Fellow at the National Air and Space Museum and is currently an Associate Fellow of the American Institute of Aeronautics and Astronautics, as well as a Fellow of the Royal Aeronautical Society. He is a Ph.D. candidate at MIT researching flight simulation technology.
David Doria holds a Ph.D. in Electrical Engineering from the Rensselaer Polytechnic Institute. During his time at the university, his research focused on LiDAR data analysis including object detection, segmentation, and 3D hole filling. After graduation, Doria spent the first few years of his career at the US Army Research Laboratory before he joined forces with HERE Technologies as a Senior Software Engineering Manager. During his time at HERE, he built and lead teams to develop, implement, and productionize algorithms for computer vision-based automatic creation of High-Definition maps for use in autonomous driving systems.
Most recently he has joined Magna International as Director of Engineering within the Autonomous Driving division. David leads Magna’s Advanced Development Center in Boston - the next step in Magna's vision to lead the way in the ever-growing market of high-tech mobility. The Center drives many initiatives from next generation Advanced Driver Assistance Systems (ADAS) to developing new applications of mobile robotics platforms.
Vecna Robotics delivers flexible material handling solutions, e.g. autonomous pallet trucks and forklifts, to warehouses and manufacturing facilities to keep pace with rapidly accelerating customer expectations despite historic labor shortages and supply chain challenges. As Chief Technical Officer, Dr. Dydek ensures that Vecna Robotics remains technology leaders in this space. Dr. Dydek’s graduate research involved the conception, design, and implementation of advanced, nonlinear controllers with applications to manned and unmanned aerial vehicles. He received the National Defense Science and Engineering Graduate Fellowship from the Department of Defense in 2006. Dr. Dydek has a BS degree in Mechanical Engineering with a minor in Control and Dynamical Systems from the California Institute of Technology and MS and PhD degrees in Mechanical Engineering from the Massachusetts Institute of Technology.
Chuchu Fan is the Wilson Assistant Professor in the Department of Aeronautics and Astronautics at MIT, where she leads the Reliable Autonomous Systems Lab (REALM). Fan’s research utilizes rigorous mathematics, including formal methods, machine learning, and control theory, for the design, analysis, and verification of safe autonomous systems. Her recent research focuses on certificate learning alongside learning-enabled robotics control systems to provide concise, data-driven proofs that guarantee safety and stability of a learned control system, and applying these tools to practical robotics problems. Fan received her PhD in computer engineering from the University of Illinois at Urbana-Champaign and BE in automation from Tsinghua University, China.
Learning-enabled data-driven methods have demonstrated impressive empirical performance on challenging autonomous systems. But this performance comes at the cost of reduced transparency and lack of guarantees on the safety or stability of the systems. In this talk, I will present several of our recent efforts that combine machine learning with formal methods and control theory to enable the design of dependable and safe autonomous systems. The approach we took, called neural certificates, provides supervision during training by allowing safety and stability requirements to influence the training process. As a result, the learned policies can achieve a much-improved performance on safety and stability, especially on complex autonomous systems with a large number of agents, following nonlinear and nonholonomic dynamics and needing to satisfy high-level specifications.
Leslie Kaelbling is a Professor at MIT. She has an undergraduate degree in Philosophy and a PhD in Computer Science from Stanford, and was previously on the faculty at Brown University. She was the founding editor-in-chief of the Journal of Machine Learning Research. Her research agenda is to make intelligent robots using methods including estimation, learning, planning, and reasoning. She is not a robot.
We, as robot engineers, have to think hard about our role in the design of robots and how it interacts with learning, both in "the factory" (that is, at engineering time) and in "the wild" (that is, when the robot is delivered to a customer). I will share some general thoughts about the strategies for robot design and then talk in detail about some work I have been involved in, both in the design of an overall architecture for an intelligent robot and in strategies for learning to integrate new skills into the repertoire of an already competent robot.
Nidhi Seethapathi is an assistant professor in the Department of Brain and Cognitive Sciences and the Department of Electrical Engineering and Computer Science (EECS) at MIT. She builds computational predictive models of human movement with applications to autonomous and robot-aided neuromotor rehabilitation. For this, she uses a combination of tools and approaches from dynamics, control theory, and machine learning. Her group works to understand the objectives governing movement decisions, strategies used to execute movement, and how new movements are learned. Seethapathi received a PhD in mechanical engineering from Ohio State University, and a BTech in mechanical engineering from the Veermata Jijabai Technological Institute.
The best current robots still fall short of the versatility, efficiency, stability, and robustness to uncertainty achieved by the human sensorimotor control system. One way to understand how such remarkable performance is achieved is to develop theoretical frameworks and computational models that capture empirical evidence of how humans select, execute, and learn movements. In this talk, I will summarize the findings of our past and ongoing work seeking to uncover the computational principles underlying efficient locomotor control in natural environments, the strategies underlying robust locomotion in the presence of intrinsic and extrinsic perturbations, and the seamless integration of sensorimotor control with higher-level cognition across different timescales. These principles can provide a blueprint for engineers seeking to develop autonomous and robot-aided systems that exhibit intelligence comparable to biological sensorimotor control.