Agenda
08:30-09:00: Registration
09:00-09:30: Welcome (Chair: Hélène Kirchner, Inria) – Antoine Petit, Inria CEO, and Costas Spanos, CITRIS Director.
09-30-11:30: Keynotes (Chair: Valérie Issarny, Inria)
- The Practical and System Challenges of Managing Big Data, Prof Amr El Abbadi, UC Santa Barbara
- Predicting how macromolecules interact: insights into challenges for kinematics and data science, Julie Bernauer, NVIDIA, Formerly Inria
11:30-12:00: Break in SDH Atrium
12:00-01:00: Panel on Big Data Science (Chairs: Deborah Agarwal, LBNL & Laura Grigori, Inria)
01:00-02:30: Lunch Break in SDH Atrium
02:00-04:00: Keynotes (Chair: Christine Morin, Inria)
- Robots with their Heads in the Cloud, Ken Goldberg, UC Berkeley
- Spatio-temporal Localization of Actions in Videos, Zaid Harchaoui, Inria
06:00: Reception at the Residence of France, 100 Edgewood Avenue, San Francisco
Keynote speeches
Associate Team BIGDATANET
The Practical and System Challenges of Managing Big Data
Big data applications have become indispensable in as diverse fields as the environment, commerce, social science and geography. The 3 “V”s (Volume, Velocity and Variety) have caused fundamental challenges to the ways traditional data management systems are designed and implemented as well as how data is consumed and analyzed. Major Privacy concerns arise when big data is managed and stored remotely in the Cloud. In this talk, we will explore some of the system and privacy challenges posed by Big Data, using a variety of different applications. We will focus on some of the challenges as data scales out in large data centers as well as in multiple data centers for fault-tolerance in the face of catastrophic failures.
Bio
Amr El Abbadi is a Professor of Computer Science at the University of California, Santa Barbara. He received his B. Eng. from Alexandria University, Egypt, and his Ph.D. from Cornell University. Prof. El Abbadi is an ACM Fellow, AAAS Fellow, and IEEE Fellow. He was Chair of the Computer Science Department at UCSB from 2007 to 2011. He has served as a journal editor for several database journals, including, currently, The VLDB Journal, IEEE Transactions on Computers and The Computer Journal. He has been Program Chair for multiple database and distributed systems conferences, most recently SIGSPATIAL GIS 2010, ACM Symposium on Cloud Computing (SoCC) 2011, COMAD (India) 2012 and the first ACM Conference on Social Networks (COSN) 2013. He currently serves on the executive committee of the IEEE Technical Committee on Data Engineering (TCDE) and was a board member of the VLDB Endowment from 2002 to 2008. In 2007, Prof. El Abbadi received the UCSB Senate Outstanding Mentorship Award for his excellence in mentoring graduate students. In 2013, his student, Sudipto Das received the SIGMOD Jim Gray Doctoral Dissertation Award. Most recently Prof. El Abbadi was the co-recipient of the Test of Time Award at EDBT/ICDT 2015. He has published over 300 articles in databases and distributed systems and has supervised over 30 PhD students
Associate Team ITSNAP
Predicting how Macromolecules Interact: Insights into Challenges for Kinematics and Data Science
The biological function of macromolecules, such as proteins and nucleic acids, relies heavily on their interactions with their partners. The prediction of how molecules interact and how they can create large assemblies acting as nanomachines is essential for our understanding of biology but also for therapeutics and nanotechnology design. Resorting to simple geometric coarse-grained modelling and machine learning strategies, such as genetic algorithms and support vector machines, we have shown that scoring the putative complex structures can be very much improved to reach the accuracy needed for experiment design and analysis, at least in a semi-rigid body context. From that proof of concept studies, most of the prediction strategies for docking now use machine learning for scoring optimization. Being able to predict the structure and the way molecular partners deform upon binding is also key to obtain better predictions, in particular for non-coding RNAs that are essential to target oncogenes. Our efforts in RNA structure prediction techniques have shown that robotics-inspired strategies, data based parameterization of energy functions and statistical techniques largely improve the accuracy of structure prediction. Our understanding of the dynamics of non-coding RNAs for biological processes, combined with clustering techniques, allows for efficient and flexible cross-docking analysis for protein-RNA complexes.
Bio
Julie Bernauer attended ENS Cachan from 2001 to 2004 where she received a degree in Physical Chemistry. She obtained her PhD from Université Paris-Sud in 2006 while performing research in the Yeast Structural Genomics group. Her thesis focused on the use of Voronoi models for modelling protein complexes. After a post-doctoral position at Stanford University with Pr. Michael Levitt, Nobel Prize in Chemistry 2013, she joined Inria, the French National Institute for Computer Science. While Senior Research Scientist at Inria, Adjunct Associate Professor of Computer Science at École Polytechnique and Visiting Research Scientist at SLAC, her work focused on computational methods for structural bioinformatics, specifically scoring functions for macromolecule docking using machine learning, and statistical potentials for molecular simulations. She was the first to successfully introduce machine learning for coarse-grained models in the CAPRI challenge. Julie Bernauer recently joined NVIDIA Corporation as Senior Solutions Architect for Machine Learning.
-
Prof. Ken Goldberg, University of California, Berkeley
Robots with their Heads in the Cloud
The next generation of robots will be more social than solitary. Rather than viewing every robot as an isolated system with limited computation and memory, roboticists are now exploring how robots and automation systems can actively exchange information and resources via networks. Google’s self-driving car exemplifies this trend. It uses networks to index maps, images, and data on prior driving trajectories, weather, and traffic to determine spatial localization and make decisions. Data from each car is shared via the network for statistical optimization and machine learning performed by grid computing in the Cloud.
Building on emerging advances in cloud computing, big data, machine learning, open-source, and the Internet of Things, this paradigm has potential to significantly increase the capabilities of robots and automation systems. This talk will describe the latest research, including superhuman surgery and cloud-based grasping, and present a short documentary film, Why We Love Robot.
Bio
Ken Goldberg is an artist and UC Berkeley professor. He and his students investigate robotics, automation, art, and social media. Goldberg directs the Automation Sciences Research Lab, co-directs the Center for Automation and Learning for Medical Robotics, and is Faculty Director of the CITRIS Data and Democracy Initiative. Goldberg earned dual degrees in Electrical Engineering and Economics from the University of Pennsylvania (1984) and MS and PhD degrees from Carnegie Mellon University (1990). He joined the UC Berkeley faculty in 1995 where he is Professor of Industrial Engineering and Operations Research (IEOR), with secondary appointments in Electrical Engineering/Computer Science (EECS), Art Practice, the School of Information, and in the Department of Radiation Oncology at the UCSF Medical School. Goldberg has published over 200 peer-reviewed technical papers on algorithms for robotics, automation, and social information filtering; his inventions have been awarded eight US Patents. He is Editor-in-Chief of the IEEE Transactions on Automation Science and Engineering (T-ASE), Co-Founder of the African Robotics Network (AFRON), Co-Founder of the Berkeley Center for New Media (BCNM), Co-Founder and CTO of Hybrid Wisdom Labs, Co-Founder of the Moxie Institute, and Founding Director of UC Berkeley’s Art, Technology, and Culture Lecture Series. Goldberg’s art installations are related to his research and have been exhibited at venues such as the Whitney Biennial, Berkeley Art Museum, SF Contemporary Jewish Museum, Pompidou Center, Buenos Aires Biennial, and the ICC in Tokyo. Goldberg co-wrote three award-winning Sundance documentary films, “The Tribe”, “Yelp”, and “Connected: An Autoblogography of Love, Death, and Technology” and co-directed the Emmy-Nominated Short Doc “Why We Love Robots.” Goldberg is represented by the Catharine Clark Gallery in San Francisco. Goldberg was awarded the Presidential Faculty Fellowship in 1995 by President Clinton, the National Science Foundation Faculty Fellowship in 1994, the Joseph Engelberger Robotics Award in 2000, and elected IEEE Fellow in 2005. Goldberg lives in the Bay Area with his daughters and wife, filmmaker and Webby Awards founder Tiffany Shlain.
Associate Team HYPERION
Spatio-temporal Localization of Actions in Videos
We propose a novel approach for localization of human actions in real-world videos, that performs detection of actions both in the spatial domain and in the temporal domain. We present promising experimental results on the challenging UCF sports dataset, and on the HighFive dataset.
Joint work with Georgia Gkioxari and Jitendra Malik (UC Berkeley), Philippe Weinzaepfel and Cordelia Schmid (Inria).
Bio
Zaid Harchaoui is a tenured permanent researcher at Inria. He received his PhD from ParisTech (Paris, France), and his PhD thesis was on regularized kernel-based methods for detection. He received NIPS reviewer award. He gave a tutorial on “Frank-Wolfe, greedy algorithms, and friends” at ICML’14, and on “Large-scale visual recognition” at CVPR’13. He recently co-organized the workshop on “Optimization for Machine Learning” at NIPS’14, and the “Optimization and Statistical Learning” workshop in 2015 and 2013 in Ecole de Physique des Houches (France).
Panel on Big Data Science
Data analytics meets high performance computing
This panel will focus on large-scale data analytics and the software and hardware environments to support data science. It will highlight emerging directions of research in this area, emphasize already existing as well as potential synergies between data analytics and high performance computing, and outline prospects for the future. The panelists include renowned experts in algorithms, data management, visualization, and data-driven applications.
Moderators
Dr Deborah Agarwal, Senior Fellow, Lawrence Berkeley National Laboratory.
Dr Laura Grigori, Senior Research Scientist, Inria.
Panelists
-
Katie Antypas, NERSC
-
Wes Bethel, LBNL
-
Loic Esteve, Inria
-
Benjamin Recht, UC Berkeley
