Doors open at 10:00AM ET.
The Columbia Center of AI Technology (CAIT) in collaboration with Amazon was founded in 2020 with a mission to better society through the development and adoption of advanced AI technologies contributing to a more secure, connected, creative, sustainable, healthy, and equitable humanity. A major goal within this mission is to create a community of scholars and practitioners at the leading edge of AI and AI technology development.
Â
The 2nd Annual CAIT Symposium will bring together researchers, leaders, and experts from academia, industry, and government to discuss advances and new challenges in the field of AI technology, specifically related to autonomy and trust.
Vice President
Amazon Alexa

Dr. Prem Natarajan is vice president for Amazon Alexa, leading a multi-disciplinary team to make Alexa a trusted AI assistant, advisor, and companion for everyone, everywhere. His team’s product, engineering, and scientific advances have improved how customers experience Alexa through advances in natural language understanding, entity linking and resolution, and related machine learning technologies. Dr. Natarajan and his team are now focused on advancing generalizable AI, combining the best of human-like intelligence with machine learning to accelerate the future of ambient intelligence – where the underlying AI seamlessly blends into your environment, connects heterogeneous services and devices, and adapts on your behalf to provide greater utility.
Â
Prior to Amazon, he was senior vice dean of engineering at the University of Southern California, where he led nationally influential DARPA- and IARPA-sponsored research efforts in biometrics, face recognition, OCR, natural language processing, media forensics, and forecasting. Prior to that, he served as executive vice president and principal scientist for speech, language, and multimedia at Raytheon BBN Technologies.
Â
Dr. Natarajan is a well-recognized thought leader in AI and a keynote speaker who is frequently interviewed by publications like CNN, MIT Tech Review, Economist and Fast Company. He is the author and co-author of more than 200 published papers in research areas such as speech recognition, machine translation and computer vision. He also holds several patents in the field of conversational AI and machine learning. He received his PhD and master’s degree in Electrical Engineering at Tufts University, and a bachelor’s degree in Engineering from Savitribai Phule Pune University in India.
Dean of Columbia Engineering and Morris A. and Alma Schapiro Professor
Columbia University
Dean Chang is the dean of Columbia Engineering, where he leads the education, research, and innovation mission. He received his BS from National Taiwan University in 1985 and his PhD from the University of California-Berkeley in 1993.
At Columbia Engineering, Dean Chang has greatly contributed to the growth and advancement of the School, propelling it to be one of the top engineering programs in the nation. He plays a key role in the creation of the School’s guiding vision, Columbia Engineering for Humanity.
Dean Chang is the Morris A. and Alma Schapiro Professor with appointments in the departments of Electrical Engineering and Computer Science. As one of the most influential experts in multimedia, computer vision and artificial intelligence, his research has led to spinoff companies and licensed technology in multimedia search. The image search tools developed by his group have been used by more than 200 law enforcement agencies in fighting online human trafficking crimes. He has launched AI tools for online disinformation detection and attribution.
Dean Chang is a Fellow of the American Association for the Advancement of Science, the Association for Computing Machinery, and IEEE, and an elected member of Academia Sinica. He received an Honorary Doctorate from the University of Amsterdam and the Great Teacher Award from the Society of Columbia Graduates. He is also the inaugural director for the Columbia Center of AI Technology in collaboration with Amazon.
Round Table Discussions (In-Person Only; Location: Alfred Lerner Hall Room 555)
 Â
  TOPICS
  Public Data Resources
  Responsible AI - Privacy
  Responsible AI - Fairness
The purpose of these in-person breakout discussions is to showcase the interesting research being done in these areas at Amazon, facilitate an exchange of research capabilities, seed connections between Columbia faculty and Amazon scientists, and fuel future research collaborations.
Fletcher Jones Foundation Endowed Chair in Computer Science and Professor of Computer Science and Electrical & Computer Engineering
University of Southern California
Big Data and Small Models: Lessons for Robotics
We have recently demonstrated the possibility of learning drone swarm controllers that are zero-shot transferable to real quadrotors via large-scale, multi-agent, end-to-end reinforcement learning. We train policies parameterized by neural networks that can control individual drones in a swarm in a fully decentralized manner. Our policies, trained in simulated environments with realistic quadrotor physics, demonstrate advanced flocking behaviors, perform aggressive maneuvers in tight formations while avoiding collisions with each other, break and re-establish formations to avoid collisions with moving obstacles, and efficiently coordinate in pursuit-evasion tasks. We will demonstrate the successful deployment of the model learned in simulation to highly resource-constrained physical quadrotors performing station-keeping and goal-swapping behaviors. Motivated by these results, and the observation that neural control of memory-constrained, agile robots requires small, yet highly performant models, we have begun a project that leverages graph hypernetworks to learn hyperpolicies trained with off-policy reinforcement learning. This results in networks that are two orders of magnitude smaller than commonly used networks yet encode policies comparable to those encoded by much larger networks trained on the same task. Our method can be appended to any off-policy reinforcement learning algorithm, without any change in hyperparameters, we illustrate this by showing early results across locomotion and manipulation tasks. The talk will conclude with some thoughts on the generality of such approaches for a more general class of devices with modest computational capabilities.
Gaurav S. Sukhatme holds the Fletcher Jones Foundation Endowed Chair in Computer Science at the University of Southern California (USC). He is Professor of Computer Science and Electrical & Computer Engineering and serves as the Executive Vice Dean at the USC Viterbi School of Engineering. He is an Amazon Scholar. He received his undergraduate education at IIT Bombay in Computer Science and Engineering, and M.S. and Ph.D. degrees in Computer Science from USC. He is the co-director of the USC Robotics Research Laboratory and the director of the USC Robotic Embedded Systems Laboratory, which he founded in 2000. His research interests are in networked robots, learning robots and field robotics. He has published extensively in these and related areas. Sukhatme has served as PI on numerous NSF, DARPA and NASA grants. He was a Co-PI on the Center for Embedded Networked Sensing (CENS), an NSF Science and Technology Center. He is a fellow of the AAAI, the IEEE and a recipient of the NSF CAREER award and the Okawa foundation research award. He is one of the founders of the Robotics: Science and Systems conference and the Editor-in-Chief of Autonomous Robots.
Senior Staff Research Scientist
Challenges and Opportunities in Large-Scale Language Modeling
In the last few years, neural network language models have been scaled to hundreds of billions of parameters, resulting in breakthrough capabilities as few-shot learners and conversational agents. In this talk, I will go over some of the challenges in training language models at this scale. Additionally, I will give an overview of key research areas in large-scale language modeling besides scale itself, such as meta-learning, optimization, sparsity, and efficient architectures.
Jacob Devlin is a Senior Staff Research Scientist at Google, where he works on deep learning models for natural language understanding. He is best known for developing the BERT model for language understanding. More recently, he co-led the PaLM project, which is Google's largest-scale language modeling effort to date.
Sergul has been a Senior Applied Scientist at Amazon AI, leading privacy in ML efforts. Her research interests are Differential Privacy and Robust Machine Learning. Before Amazon, Sergul was an Assistant Professor at the Stevens Institute of Technology, NJ, USA. She is also the General Chair for Women in ML Workshop at NeurIPS’22. Sergul received her PhD degree from the Signal and Image Processing Institute at University of Southern California in 2014. Her PhD work was on developing robust connectivity measures for neuroimaging data. Sergul’s scientific work has been published in AI and medical imaging venues.
Differentially Private Synthetic Data
In this talk, I will describe a formal solution for privacy-preserving data exchange. Imagine a hospital that wants to share its data with external researchers to accelerate scientific findings. Due to the sensitive nature of patient data, the data is anonymized to protect the privacy of individuals. However, anonymized data does not provide formal privacy guarantees and the utility of the data is likely to be destroyed. Instead, with help of machine learning tools, a synthetic version of the original data can be used for data exchange. The synthetic data represents the original data but is not the exact replica. To provide privacy guarantees, we rely on the notion of differential privacy. The differentially private synthetic data can then be shared between parties with formal privacy guarantees. I will present our approach for generating private synthetic data and demonstrate the usefulness for query release and machine learning tasks.
Â
Suman Jana is an associate professor in the department of computer science and the data science institute at Columbia University. His primary research interest is at the intersections of computer security and machine learning. His research has received six best paper awards, a CACM research highlight, a Google faculty fellowship, a JPMorgan Chase Faculty Research Award, an NSF CAREER award, and an ARO young investigator award.
Efficient Neural Network Verification using Branch and Bound
In this talk, I will describe two recent Branch and Bound (BaB) verifiers developed by our group to ensure different safety properties of neural networks. The BaB verifiers involve two main steps: (1) recursively splitting the original verification problem into easier independent subproblems by splitting input or hidden neurons; and (2) for each split subproblem, using fast but incomplete bound propagation techniques to compute sound estimated bounds for the outputs of the target neural network. One of the key limitations of existing BaB verifiers is computing tight relaxations of activation functions' (i.e., ReLU) nonlinearities. Our recent works (α-CROWN and β-CROWN) introduce a primal-dual approach and jointly optimize the corresponding Lagrangian multipliers for each ReLU with gradient ascent. Such an approach is highly parallelizable and avoids calls to expensive LP solvers. Our verifiers not only provide tighter output estimations than existing bound propagation methods but also can fully leverage GPUs with massive parallelization. Our verifier, α, β-CROWN (alpha-beta-CROWN), won the second International Verification of Neural Networks Competition (VNN-COMP 2021) with the highest total score.
Govind Thattai is a Principal Scientist at Alexa AI, and is leading the Multimodal science efforts for Embodied AI and Visual Question Answering. Prior to joining Amazon in 2018, Govind has worked in the areas of Speech Recognition, Computer Vision and NLU at BBN, KLA and eBay.Â
Embodied AI at AlexaÂ
Embodied AI is the AI that enables a robotic agent to perceive, navigate, interact and learn in a 3D physical environment. This talk describes Alexa’s initiative to present Embodied AI as a University Challenge for Alexa Prize, and the new Embodied AI framework called Arena.
Ali Hirsa is a professor, director of financial engineering program, and director of Center for AI in business analytics & FinTech at Columbia University. Ali has worked at both sell-side and buy-side for more than 25 years. Ali’s research interests are AI/ML/DL applications in asset management and finance. Ali is author of two textbooks and is Editor-in-Chief of Journal of Investment Strategies. He is a frequent speaker at academic and practitioner conferences. Ali received his PhD in Applied Mathematics from University of Maryland at College Park under the supervision of Professors Howard C. Elman and Dilip B. Madan.
Eshan Bhatnagar is Head of Product at Alexa AI Natural Understanding team where he leads a multidisciplinary product, design and user research organization with a focus on applying AI across Natural Language Understanding, Computer Vision, Multimodal Learning & Generation, and Robotics to deliver delightful experiences to millions of Amazon customers worldwide. Eshan was an early product leader in the Amazon Alexa organization where he helped build the original Echo Show, and founded new product areas from scratch that are used by millions of customers worldwide. Eshan joined Amazon in 2013 where he led Commerce, Payments and Currency products at AWS, and helped launch AWS's local business in India. Prior to Amazon, Eshan led R&D efforts on multicore application processors at Freescale Semiconductors, and conducted technology and financial due-diligence for early-stage startups at Amiti Ventures. Eshan holds a bachelor’s degree in Electronics and Communications Engineering from Manipal Institute of Technology, India, and an MBA in Finance, Accounting, Entrepreneurship, and Marketing from the University of Chicago Booth School of Business.
David Boothe was born in Washington DC in 1967. He received a B.A. in Philosophy from the University of Maryland, College Park in 1989, and a Ph.D. in Computational Neuroscience from the Neuroscience and Cognitive Sciences program University of Maryland in 2007.  He did Post-Doctoral research at the Rehabilitation Institute of Chicago, and the Feinberg School of Medicine at Northwestern University.  He is currently the program manager of the Strengthening Teamwork for Novel Groups Collaborative Research Alliance at the Combat Capabilities Development Command, Army Research Laboratory.  He specializes in human machine teaming, neuro-inspired intelligent system design, and brain modeling and simulation.
Dr. Kaleb McDowell has a B.S in operations research and industrial engineering from Cornell University, and an M.S. in kinesiology and a Ph.D. in neuroscience and cognitive science both from the University of Maryland, College Park. He currently leads ARL’s Human-Guided System Adaptation Research Branch. Since joining ARL in 2003, Dr. McDowell has led several major research and development programs focused on neuroscience and human-intelligent technology integration; served as the chief scientist of the Humans in Complex Systems Directorate, and receiving Army Research and Development Achievement awards in 2007, 2009, and 2013; and ARL’s Director’s Awards in 2020.
Jeannette M. Wing is the Executive Vice President for Research and Professor of Computer Science at Columbia University. She joined Columbia in 2017 as the inaugural Avanessians Director of the Data Science Institute. From 2013 to 2017, she was a Corporate Vice President of Microsoft Research. She is Adjunct Professor of Computer Science at Carnegie Mellon where she twice served as the Head of the Computer Science Department and had been on the faculty since 1985. From 2007-2010 she was the Assistant Director of the Computer and Information Science and Engineering Directorate at the National Science Foundation. She received her S.B., S.M., and Ph.D. degrees in Computer Science, all from the Massachusetts Institute of Technology.
Tang Professor of Operations and Senior Vice Dean for Research and Academic Programs, Columbia University
Garud Iyengar is the Tang Professor of Operations at Columbia Engineering. He received his B. Tech. in Electrical Engineering from IIT Kanpur, and an MS and PhD in Electrical Engineering from Stanford University. His research interests are broadly in control, machine learning and optimization. His published works span a diverse range of fields, including information theory, applied mathematics, operations research, economics and financing engineering. His current projects focus on the areas of large-scale power systems and supply chains, causal inference, and modeling of cellular processes. He was elected an INFORMS Fellow in 2018. He was the Chair of the Department of Industrial Engineering and Operations Research from 2013-19, and the Associate Director for Research at the Columbia Data Science Institute from 2017-19. He has been an Amazon Scholar since 2019. He is currently the Senior Vice Dean for Research and Academic Programs at Columbia Engineering.
Jack FitzGerald is a senior applied scientist with Amazon Alexa AI’s Natural Understanding group. His research interests include large language models, multilingual language understanding, efficient distributed training, and multitask modeling. Jack has been with Amazon since 2015 and Alexa since 2017. Before Amazon, he was an officer and nuclear physicist in the US Air Force.
Maryam Zolnoori, PhD, Research Scientist, Columbia University Irving Medical Center
ROUND TABLE: Public Data Resources
Maryam Zolnoori, PhD, is a research scientist at Columbia University Irving Medical Center. She received her PhD in Biomedical Informatics and her Master in both Information Technology and Health Informatics. Her research have focused on utilizing advanced data science methods and routinely generated data in clinical settings to build decision support tools for improving quality of healthcare. She is the author of more than 50 articles and recipient of several research awards from National Library of Medicine, National Institute on Aging, Federal Drug Administration, Mayo Clinic, and Columbia Center of AI Technology.
Â
Sergul has been a Senior Applied Scientist at Amazon AI, leading privacy in ML efforts. Her research interests are Differential Privacy and Robust Machine Learning. Before Amazon, Sergul was an Assistant Professor at the Stevens Institute of Technology, NJ, USA. She is also the General Chair for Women in ML Workshop at NeurIPS’22. Sergul received her PhD degree from the Signal and Image Processing Institute at University of Southern California in 2014. Her PhD work was on developing robust connectivity measures for neuroimaging data. Sergul’s scientific work has been published in AI and medical imaging venues.
Roxana Geambasu, PhD
Associate Professor of Computer Science, Columbia University
ROUND TABLE: Responsible AI - Privacy
Roxana Geambasu is an Associate Professor of Computer Science at Columbia University and a member of Columbia's Data Sciences Institute. She joined Columbia in Fall 2011 after finishing her Ph.D. at the University of Washington. For her work in cloud and mobile data privacy, she received an Alfred P. Sloan Faculty Fellowship, a Microsoft Research Faculty Fellowship, an NSF CAREER award, a ``Brilliant 10'' Popular Science nomination, an Early Career Award in Cybersecurity from the University of Washington Center for Academic Excellence, the Honorable Mention for the 2013 inaugural Dennis M. Ritchie Doctoral Dissertation Award, a William Chan Dissertation Award, two best paper awards at top systems conferences, and the first Google Ph.D. Fellowship in Cloud Computing.
Rahul Gupta is a Senior Applied Science manager at the Spoken Language Understanding Innovations (SLU-Innovations) team in Cambridge, Massachusetts. Since joining the Alexa organization, he has focused on designing NLU models for scalability and speed. Some of his more recent research focuses on Trustworthy Machine Learning with emphasis on privacy preserving techniques, fairness and federated learning. He received his PhD from the University of Southern California in 2016 on interpreting non-verbal communications in human interaction. He has published several papers at avenues such as EMNLP, ACL, NAACL, ACM Facct and IEEE-Transactions. He is also co-inventor on ten patent pending technologies at Amazon.
Shih-Fu Chang, PhD
Dean of Columbia Engineering and Morris A. and Alma Schapiro Professor
ROUND TABLE: Responsible AI - Fairness
Shih-Fu Chang is Dean of Columbia Engineering and Morris A. and Alma Schapiro Professor. He leads the education, research, and innovation mission of the School and has greatly contributed to its growth and advancement, propelling it to be one of the top engineering programs in the nation.
As one of the most influential experts in multimedia, computer vision and artificial intelligence, his research has led to development of innovative image search tools, which have been used by major media companies and law enforcement agencies in fighting online human trafficking crimes. He has also launched AI tools for online disinformation detection and attribution.
Dean Chang is a Fellow of the American Association for the Advancement of Science, the Association for Computing Machinery, and IEEE, and an elected member of Academia Sinica. He received the Great Teacher Award from the Society of Columbia Graduates and is director of the Columbia Center of AI Technology in collaboration with Amazon.
Doors open at 10:00AM ET.
The Columbia Center of AI Technology (CAIT) in collaboration with Amazon was founded in 2020 with a mission to better society through the development and adoption of advanced AI technologies contributing to a more secure, connected, creative, sustainable, healthy, and equitable humanity. A major goal within this mission is to create a community of scholars and practitioners at the leading edge of AI and AI technology development.
Â
The 2nd Annual CAIT Symposium will bring together researchers, leaders, and experts from academia, industry, and government to discuss advances and new challenges in the field of AI technology, specifically related to autonomy and trust.
The mission of the Columbia University Center of Artificial Intelligence Technology in collaboration with Amazon is to better society through the development and adoption of advanced AI technology contributing to a more secure, connected, creative, sustainable, healthy and equitable humanity.
