Intelligent Machines Symposium
Total Page:16
File Type:pdf, Size:1020Kb
MACHINE LEARNING IN RESEARCH AND APPLICATIONS CALL FOR POSTERS AND DEMONSTRATIONS INTELLIGENT MACHINES SYMPOSIUM 17 NOVEMBER 2010 “DE VEREENIGING” IN NIJMEGEN Organisation Prof. dr. H.J. Kappen | Radboud Universiteit Nijmegen / SNN A. Wanders | SNN SNN SNN is a non-for-profit organization that aims to stimulate research and applications on machine learning, Bayesian inference and neural networks in the Netherlands. SNN facilitates a network of researchers, organizations and companies in this field. Current participants are: Prof. Dr. H.J. Kappen, dr. W. Wiegerinck | Radboud Universiteit Nijmegen / SNN Prof. Dr. T. Heskes | Radboud Universiteit Nijmegen Prof. Dr. H.J. van den Herik | Tilburg center for Cognition and Communication Prof. Dr. E. Postma | Tilburg center for Cognition and Communication Prof. Dr. A. Van den Bosch | Tilburg center for Cognition and Communication Prof. Dr. J.N. Kok | Universiteit Leiden Prof. P. Grünwald | Universiteit Leiden Prof. Dr. R. Gill | Universiteit Leiden Prof. Dr.L. Schomaker, dr. M. Wiering | Rijksuniversiteit Groningen Prof. Dr. G. Weiss, dr. K. Tuyls | Universiteit Maastricht Prof. Dr. F. Groen, dr. B. Krose, dr. S. Whiteson | Universiteit van Amsterdam Prof. Dr. H. La Poutre, dr. S. Bohte | Centrum voor Wiskunde en Informatica Dr. K. Nieuwenhuis, dr. G. Pavlin | Thales Research Prof. Dr. M. Reijnders, dr. B. Duin, dr. M. Loog | Technische Universiteit Delft Prof. Dr. A. van der Vaart | Vrije Universiteit Amsterdam Intelligent Machines Can machines think? This has been a conundrum for philosophers for years, but the answer to this question also has real social importance. Modern robots can assist us in our homes and have human-like qualities. The internet provides us with personalized tools that learn from our behavior. It is therefore of more than academic importance that we learn to think clearly about the actual cognitive powers of computers, and what we can expect of them in the future. On November 17, SNN organizes a one-day symposium in Nijmegen, entitled "Intelligent Machines". During the day, invited research leaders in the field of machine learning and robotics will present examples of intelligence in computers and will discuss their views for the future. In addition, posters will provide a comprehensive overview of the research activities on machine learning in the Netherlands. Based on our experience with similar events in the past, we expect approximately 40 poster presentations. There is also an opportunity for companies to present their research and products. The purpose of this meeting is to provide a platform for discussion and interaction between the academic and industrial research and development communities in the Netherlands. For Whom The symposium is designed to address a broad audience. The plenary talks aim at anyone who is interested in how intelligent computers may affect our society, now and in the future. The poster sessions present more technical results and aim provide concrete examples of how machine learning research is used in numerous applications. 1 2 What is machine learning? Machine learning is a scientific discipline that is concerned with the design and development of algorithms that allow computers to learn behaviors based on empirical data, such as from sensor data or databases. A major focus of machine learning research is to automatically learn to recognize complex patterns and make intelligent decisions based on data; the difficulty lies in the fact that the set of all possible behaviors given all possible inputs is too complex to describe generally in programming languages, so that in effect programs must automatically describe programs. Machine learning methods are at the basis of many applications, ranging from vision to language processing, forecasting, pattern recognition, games, data mining, expert systems and robotics. The modern field of machine learning integrates many distinct approaches such as probability theory, logic, combinatoric optimization, search, statistics, reinforcement learning and control theory. Where and When The meeting will take place on Wednesday the 17 th of November 2010 from 9:30 a.m. until 6:00 p.m. in the Concert building “De Vereeniging” in Nijmegen. This symposium has no registration or attendance costs and includes a free lunch. “De Vereeniging” Nijmegen | Keizer Karelplein 2d | 6511 NC Nijmegen | tel +31 (0)24 3608135 | SNN + 31 24 3614245 | mobile SNN 06 2817 6686 There are some parkingplaces available. It’s about 5 minutes walking from Central Train Station Nijmegen. 3 Program 09:30 Registration 09:50 Opening Prof. Bert Kappen | Radboud University Nijmegen, SNN 10:00 Tutorial: Bridging the gap between machines and people I Prof. Nicholas Roy | Massachusetts Inst. of Technology | Cambridge | USA 10:45 Coffee 11:00 Tutorial: Bridging the gap between machines and people II Prof. Nicholas Roy | Massachusetts Inst. of Technology | Cambridge | USA 11:45 Lunch and Posters There will be a lunch buffet held across the different meeting rooms. During lunch, stands and poster presentations can be mounted for display. 13:00 Machine Learning at Yahoo! Dr. Kishore Papineni | Yahoo! Research | New York | USA 13:45 Learning to get smart Prof. Edgar Körner | Honda Research Institute Europe | Main | Germany 14:30 Poster session 15:45 What is Intelligence? Prof. Nello Cristianini | University of Bristol | Bristol | UK 16:30 Panel discussion moderated by Bas Haring 17.00 Drinks and Posters 18:00 Closing 10:00 Machine Learning and Robotics: Bridging the gap between machines and people Nicholas Roy from Massachusetts Institute of Technology In the last few years, how robots operate in the world has advanced considerably. Examples include the autonomous vehicles in the DARPA Grand Challenges and Urban Challenge, the considerable work in robot mapping, and the growing interest in home and service robots. However, these example technologies and systems are still mostly restricted to research prototypes. One obstacle to getting more widely useful robots is that the way robots reason about their world is still pretty different to how people reason. Robots think in terms of point features, dense occupancy grids and action cost maps. People think in terms of landmarks, segmented objects and tasks (among other representations). There are good reasons why these are different, and robots are unlikely to ever reason about the world in the same way that people do. But, there has been recent work in bridging the gap between low-level geometry and control, and higher-level semantic representations. Nicholas Roy is an Associate Professor in the Department of Aeronautics & Astronautics at the Massachusetts Institute of Technology and a member of the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT. He received his Ph.D. in Robotics from Carnegie Mellon University in 2003. His research interests include autonomous systems, micro air vehicles, mobile robotics, human-computer interaction, decision- making under uncertainty and machine learning. 11:45 Lunch and Posters There will be a lunch buffet held across the different meeting rooms. During lunch, stands and poster presentations can be mounted for display. 4 5 13:00 Machine Learning at Yahoo! Kishore Papineni, Yahoo! Research New York As a large internet portal with diverse audience, Yahoo! is constantly faced with choosing the most appropriate object (a news story, a set of urls, an advertise- ment) for a specific user in context. It must choose the object(s) out of a large number of choices (millions to billions) in a short time (20-50 milliseconds). Extreme personalization means modeling the user’s needs at that moment well. Serving systems must adapt to changes in the environment — e.g shifts in advertising marketplace, drifts in user’s needs and tastes, hot news stories becoming cold, etc. The underlying learning machinery must sometimes learn in an adversarial setting (eg email spam detection). While the training data is web- scale that grows terabytes a day, it is still sparse because only a tiny fraction of all possible combinations of webpages, users, and advertisements will be seen in the historical data. A rich combination of mathematics, statistics, computer science, and economics forms the basis of many serving systems at Yahoo. This talk gives an overview of some of the learning/optimization challenges and examines one aspect in depth. Kishore Papineni graduated with a PhD in Electrical Engineering specializing in feedback control theory from Rice University in 1995. From 1995 to 2006, he was a Research Staff Member at the IBM T.J. Watson Research Center, Yorktown Heights, New York. During this period, he worked on various natural language processing technologies such as natural language understanding, dialog management and statistical machine translation, managing IBM's SMT department from 2001-2006. He was a founding Editor-in-Chief of ACM Transactions on Speech and Language Processing from 2003-2007. In 2006, he joined Yahoo! Research where he is the head of Machine Learning. His interests include optimization, estimation, control, and computational advertising. 6 13:45 Learning to get smart –how to cope with the challenge of interaction within a natural environment Edgar Körner Honda Research Institute Europe GmbH Progress in machine learning has enabled technical artifacts to perform quite well for well specified tasks for which the necessary knowledge can be either provided by preprogramming a rule base or supervised learning. How far we are still