Invited Speakers
Christian Theobalt is a Professor of Computer Science and the head of the research group "Graphics, Vision, & Video" at the Max-Planck-Institute (MPI) for Informatics, Saarbrücken, Germany. He is also a Professor of Computer Science at Saarland University, Germany. From 2007 until 2009 he was a Visiting Assistant Professor in the Department of Computer Science at Stanford University. He received his MSc degree in Artificial Intelligence from the University of Edinburgh, his Diplom (MS) degree in Computer Science from Saarland University, and his PhD (Dr.-Ing.) from the Max-Planck-Institute for Informatics.
In his research he looks at algorithmic problems that lie at the intersection of Computer Graphics, Computer Vision and machine learning, such as: static and dynamic 3D scene reconstruction, marker-less motion and performance capture, virtual and augmented reality, computer animation, appearance and reflectance modelling, intrinsic video and inverse rendering, machine learning for graphics and vision, new sensors for 3D acquisition, advanced video processing, as well as image- and physically-based rendering. He is also interested in using reconstruction techniques for human computer interaction. For his work, he received several awards, including the Otto Hahn Medal of the Max-Planck Society in 2007, the EUROGRAPHICS Young Researcher Award in 2009, the German Pattern Recognition Award 2012, and the Karl Heinz Beckurts Award in 2017. He received two ERC grants, an ERC Starting Grant in 2013 and an ERC Consolidator Grant in 2017. In 2015, he was elected as one of the top 40 innovation leaders under 40 in Germany by the business magazine Capital. Christian Theobalt is also a co-founder of an award-winning spin-off company from his group - www.thecaptury.com - that is commercializing one of the most advanced solutions for marker-less motion capture.
To see the full presentation video click here
Timnit Gebru is currently a research scientist at Google in the ethical AI team. Prior to that she did a postdoc at Microsoft Research, New York City in the FATE (Fairness Transparency Accountability and Ethics in AI) group, where she studied algorithmic bias and the ethical implications underlying any data mining project (see this New York Times article for an example of my work). Dr. Gebru received her PhD from the Stanford Artificial Intelligence Laboratory, studying computer vision under Fei-Fei Li. Her thesis pertains to data mining large scale publicly available images to gain sociological insight, and working on computer vision problems that arise as a result. The Economist, The New York Times and others have covered part of this work. Some of the computer vision areas she is interested in include fine-grained image recognition, scalable annotation of images, and domain adaptation. Prior to joining Fei-Fei's lab she worked at Apple designing circuits and signal processing algorithms for various Apple products including the first iPad. She also spent an obligatory year as an entrepreneur (as all Stanford undergrads seem to do). Her research was supported by the NSF foundation GRFP fellowship and the Stanford DARE fellowship
Manolis Savva is an Assistant Professor in the School of Computing Science at Simon Fraser University in Vancouver, Canada and a visiting researcher at Facebook AI Research. His research focuses on analysis, organization and generation of 3D content through a human-centric lens of "common sense" semantics. The methods that he works on are stepping stones towards a holistic form of 3D scene understanding revolving around people, with applications in computer graphics, computer vision, and robotics.
Dr. Savva worked as a postdoctoral research associate at the Princeton University Computer Graphics and Vision Lab. He received his Ph.D. from Stanford University, under the supervision of Pat Hanrahan. Prior to that, he received his undergraduate degree in Physics and Computer Science at Cornell University.
His research interests span Human-centric 3D scene analysis, 3D scene synthesis for VR/AR content creation, and learning through simulation, Connecting natural language with 3D representations and Data visualization
Hima Lakkaraju is an Assistant Professor at Harvard University since January 2020 where She was a postdoctoral fellow earlier. She has recently graduated with a PhD in Computer Science from Stanford University. Her research focuses on building accurate, interpretable, and fair AI models which can assist decisions makers (e.g., judges, doctors) in critical decisions (e.g., bail decisions). Her work finds applications in high-stakes settings such as criminal justice, healthcare, public policy, and education. At the core of her research lie rigorous computational techniques spanning AI, ML, and econometrics. Hima has recently been named one of the 35 innovators under 35 by MIT Tech Review, one of the innovators to watch by Vanity Fair, and has received several fellowships and awards including the Robert Bosch Stanford graduate fellowship, Microsoft research dissertation grant, Google Anita Borg scholarship, IBM eminence and excellence award, and best paper awards at SIAM International Conference on Data Mining (SDM) and INFORMS.
Pushmeet Kohli leads the research for the Science, Robustness and Reliability teams at DeepMind. His research revolves around the development and deployment of robust AI systems for making progress on key challenges in the Natural Sciences, and for solving impactful real world problems. His current research interests include Safe, Reliable and Trustworthy AI, Protein Engineering, Material Science, Program Synthesis and Induction, and learning of interpretable representations of data. Pushmeet’s papers have appeared in conferences in the field of machine learning, computer vision, game theory and human computer interaction and have won awards in CVPR, WWW, CHI, ECCV, ISMAR, TVX and ICVGIP. His research has also been the subject of a number of articles in popular media outlets such as Forbes, Wired, BBC, New Scientist and MIT Technology Review. Pushmeet has served on the editorial boards of PAMI, JMLR, IJCV and CVIU.
Professor Alex 'Sandy' Pentland directs MIT Connection Science, an MIT-wide initiative, and previously helped create and direct the MIT Media Lab and the Media Lab Asia in India. He is one of the most-cited computational scientists in the world, and Forbes recently declared him one of the "7 most powerful data scientists in the world" along with Google founders and the Chief Technical Officer of the United States. He is on the Board of the UN Foundations' Global Partnership for Sustainable Development Data, co-led the World Economic Forum discussion in Davos that led to the EU privacy regulation GDPR, and was central in forging the transparency and accountability mechanisms in the UN's Sustainable Development Goals. He has received numerous awards and prizes such as the McKinsey Award from Harvard Business Review, the 40th Anniversary of the Internet from DARPA, and the Brandeis Award for work in privacy.
He is a member of advisory boards for the UN Secretary General and the UN Foundation, and the American Bar Association, and previously for Google, AT&T, and Nissan. He is a serial entrepreneur who has co-founded more than a dozen companies including social enterprises such as the Harvard-ODI-MIT DataPop Alliance . He is a member of the U.S. National Academy of Engineering and leader within the World Economic Forum.
Over the years Sandy has advised more than 70 PhD students. Almost half are now tenured faculty at leading institutions, with another one-quarter leading industry research groups and a final quarter founders of their own companies. Together Sandy and his students have pioneered computational social science, organizational engineering, wearable computing (Google Glass), image understanding, and modern biometrics. His most recent books are Social Physics, published by Penguin Press, and Honest Signals, published by MIT Press.
Interesting experiences include dining with British Royalty and the President of India, staging fashion shows in Paris, Tokyo, and New York, and developing a method for counting beavers from space.
Hany Farid is a Professor at the University of California, Berkeley with a joint appointment in Electrical Engineering & Computer Science and the School of Information. His research focuses on digital forensics, image analysis, and human perception. He received his undergraduate degree in Computer Science and Applied Mathematics from the University of Rochester in 1989, his M.S. in Computer Science from SUNY Albany, and his Ph.D. in Computer Science from the University of Pennsylvania in 1997. Following a two-year post-doctoral fellowship in Brain and Cognitive Sciences at MIT, he joined the faculty at Dartmouth College in 1999 where he remained until 2019. He is the recipient of an Alfred P. Sloan Fellowship, a John Simon Guggenheim Fellowship, and is a Fellow of the National Academy of Inventors.
Detecting Deep-Fake Videos from Appearance and Behavior
Synthetically-generated audios and videos -- so-called deep fakes -- continue to capture the imagination of the computer-graphics and computer-vision communities. At the same time, the democratization of access to technology that can create sophisticated manipulated video of anybody saying anything continues to be of concern because of its power to disrupt democratic elections, commit small to large-scale fraud, fuel dis-information campaigns, and create non-consensual pornography. I will describe a biometric-based forensic technique for detecting face-swap deep fakes. This technique combines a static biometric based on facial recognition with a temporal, behavioral biometric based on facial expressions and head movements, where the behavioral embedding is learned using a CNN with a metric-learning objective function. I will show the efficacy of this approach across several large-scale video datasets, as well as in-the-wild deep fakes.
Tsung-Yi Lin received his PhD at Cornell NYC Tech supervised by Serge Belongie in 2017. Before h came to NYC, he spent two wonderful years at UCSD. His research interests are in Computer Vision, in particular learning visual representations for cross-view image matching and object detection. Luckily, he had the great opportunity to work with James Hays on cross-view image geolocalization and Piotr Dollár on object detection. Currently, Dr. Lin is a research scientist at Google Brain
Fisher Yu is a postdoctoral researcher at the University of California, Berkeley, working with Trevor Darrell. He pursued his Ph.D. degree at Princeton University, advised by Thomas Funkhouser. During his Ph.D. study, he also collaborated extensively with Vladlen Koltun at Intel. He obtained his bachelor degree from the University of Michigan, Ann Arbor. His research interest lies in computer vision systems, including unified representation frameworks, 3D dynamic scene analysis, and software system for interactive data processing. He will join the ETH Zurich faculty in Switzerland as an Assistant Professor in Computer Vision in late 2020.
Debadeepta Dey is a principal researcher in the Adaptive Systems and Interaction (ASI) group led by Dr. Eric Horvitz at Microsoft Research, Redmond.
Dr. Dey finished his PhD at the Robotics Institute, Carnegie Mellon University, Pittsburgh, USA, where he was advised by Prof. Drew Bagnell. He does fundamental as well as applied research in machine learning, control and computer vision with applications to autonomous agents in general and robotics in particular.
He is interested in bridging the gap between perception and planning for autonomous ground and aerial vehicles. His interests include decision-making under uncertainty, reinforcement learning, artificial intelligence and machine learning. Nowadays he is especially interested in all kinds of aerial vehicle autonomy ranging from small quadrotors to large gliders.
Dr. Dey graduated in 2007 from Delhi College of Engineering with a Bachelors in Electrical Engineering. From 2007 to 2010 he was a researcher at the Field Robotics Center, Robotics Institute, Carnegie Mellon University.
Dr. Matthias Nießner is a Professor at the Technical University of Munich where he leads the Visual Computing Lab. Before, he was a Visiting Assistant Professor at Stanford University. Prof. Nießner’s research lies at the intersection of computer vision, graphics, and machine learning, where he is particularly interested in cutting-edge techniques for 3D reconstruction, semantic 3D scene understanding, video editing, and AI-driven video synthesis. In total, he has published over 70 academic publications, including 22 papers at the prestigious ACM Transactions on Graphics (SIGGRAPH / SIGGRAPH Asia) journal and 18 works at the leading vision conferences (CVPR, ECCV, ICCV); several of these works won best paper awards, including at SIGCHI’14, HPG’15, SPG’18, and the SIGGRAPH’16 Emerging Technologies Award for the best Live Demo.
Prof. Nießner’s work enjoys wide media coverage, with many articles featured in main-stream media including the New York Times, Wall Street Journal, Spiegel, MIT Technological Review, and many more, and his was work led to several TV appearances such as on Jimmy Kimmel Live, where Prof. Nießner demonstrated the popular Face2Face technique; Prof. Nießner’s academic Youtube channel currently has over 5 million views.
For his work, Prof. Nießner received several awards: he is a TUM-IAS Rudolph Moessbauer Fellow (2017 – ongoing), he won the Google Faculty Award for Machine Perception (2017), the Nvidia Professor Partnership Award (2018), as well as the prestigious ERC Starting Grant 2018 which comes with 1.500.000 Euro in research funding; in 2019, he received the Eurographics Young Researcher Award honoring the best upcoming graphics researcher in Europe.
In addition to his academic impact, Prof. Nießner is a co-founder and director of Synthesia Inc., a brand-new startup backed by Marc Cuban, whose aim is to empower storytellers with cutting-edge AI-driven video synthesis.
Dr. Aleksandra (Saška) Mojsilović is a scientist, Head of AI Foundations at IBM Research, Co-Director of IBM Science for Social Good, and IBM Fellow. She is a Fellow of the IEEE and a member of the IBM Academy of Technology.
Saška received her PhD in Electrical Engineering in 1997 from the University of Belgrade, Belgrade, Serbia. She has worked at Bell Laboratories (1998-2000) and IBM Research (2000-present). Her main research interests include multidimensional signal processing, data science, machine learning and AI. Over the last 20 years, Saška has applied her skills to problems in computer vision, healthcare, multimedia, finance, HR, economics, social good, and AI ethics.
Dr. Tal Hassner received his M.Sc. and Ph.D. degrees in applied mathematics and computer science from the Weizmann Institute of Science in 2002 and 2006, respectively. In 2008 he joined the Department of Math. and Computer Science at The Open Univ. of Israel where he is was a Associate Professor until 2018. From 2015 to 2018, he was a senior computer scientist at the Information Sciences Institute (ISI) and a Visiting Research Associate Professor at the Institute for Robotics and Intelligent Systems, Viterbi School of Engineering, both at USC, CA, USA, working on the IARPA Janus face recognition project. From 2018 to 2019, he was a Principal Applied Scientist at AWS where he led the design and development of the latest AWS face recognition pipelines. Since 2019 he is an Applied Research Lead at Facebook AI, supporting both the text and people photo understanding teams.
How biased is my dataset? Reasoning about dataset bias with task transferability
A dataset used to train a face recognition system (a “classification task”) would be biased against people wearing eyeglasses (a “protected group”), if it contains too little inter-class variability of faces with eyeglasses. A recognition system trained on this data could then be ill-prepared to recognize faces with eyewear. Of course, this is not the only way bias can manifest. In the same example, the dataset could also be considered biased if there was insufficient intra-subject variability of eyeglasses (i.e., subjects either always wear glasses or they never do). In this case, the recognition system would wrongly consider eyewear an identity specific facial feature. In my talk, I will propose a simple new method for reasoning about such biases. Similar to others, I build on the relationship between dataset bias and the problem of transferability: measuring how well machine learning solutions trained for one task (e.g., face recognition) generalize to new tasks (e.g., eyewear / no-eyewear). I will share our surprising recent result: In many practical use cases, only the label distributions of the two classes are required in order to predict, a-priori, the success of transferring a machine learning system from one task to the other. I will then show how our novel transferability measure can be used to expose dataset biases and offer actionable insights on dataset design.
* This talk represents work done prior to joining Facebook, in collaboration with Anh Tran and Cuong Nguyen.