Keynotes

Yoshua Bengio
Canada Research Chair in Statistical Learning Algorithms.
Department of Computer
Science and Operations Research.
Université de Montréal,
Québec, Canada.
www.iro.umontreal.ca/~bengioy

Deep Learning and Cultural Evolution

Abstract:
We propose a theory and its first experimental tests, relating difficulty of learning in deep architectures to culture and language. The theory is articulated around the following hypotheses: learning in an individual human brain is hampered by the presence of effective local minima, particularly when it comes to learning higher-level abstractions, which are represented by the composition of many levels of representation, i.e., by deep architectures; a human brain can learn such high-level abstractions if guided by the signals produced by other humans, which act as hints for intermediate and higher-level abstractions; language and the recombination and optimization of mental concepts provide an efficient evolutionary recombination operator for this purpose. The theory is grounded in experimental observations of the difficulties of training deep artificial neural networks and an empirical test of the hypothesis regarding the need for guidance of intermediate concepts is demonstrated. This is done through a learning task on which all the tested machine learning algorithms failed, unless provided with hints about intermediate-level abstractions.

Biosketch:

Yoshua Bengio received a PhD in Computer Science from McGill University, Canada in 1991. After two post-doctoral years, one at M.I.T. with Michael Jordan and one at AT&T Bell Laboratories with Yann LeCun and Vladimir Vapnik, he became professor at the Department of Computer Science and Operations Research at Université de Montréal. He is the author of two books and around 200 publications, the most cited being in the areas of deep learning, recurrent neural networks, probabilistic learning algorithms, natural language processing and manifold learning. He is among the most cited Canadian computer scientists and is or has been associate editor of the top journals in machine learning and neural networks. Since 2000 he holds a Canada Research Chair in Statistical Learning Algorithms, since 2006 an NSERC Industrial Chair, and since 2005 he is a Fellow of the Canadian Institute for Advanced Research. He is on the board of the NIPS foundation and has been program chair and general chair for NIPS. He has co-organized the Learning Workshop for 14 years and co-created the new International Conference on Learning Representations. His current interests are centered around a quest for AI through machine learning, and include fundamental questions on deep learning and representation learning, the geometry of generalization in high-dimensional spaces, manifold learning, biologically inspired learning algorithms, and challenging applications of statistical machine learning. In October 2013, Google Scholar finds more than 14500 citations to his work, yielding an h-index of 52.

 

Dario Floreano
Laboratory of Intelligent Systems, EPFL, Switzerland. lis.epfl.ch
Swiss National Center of Competence in Robotics. www.nccr-robotics.ch

Bridging Natural and Artificial Evolution

Abstract:
In this talk I will show how artificial evolution can be used to address biological questions and explain phenomena for which there is no fossil record or no experimental evidence, such evolution of behavior, altruism, and communication. I will give examples related to insects and plants. Central to this endeavor is how selection mechanisms are applied and interpreted. I will also show how selection pressure can be lifted in artificial evolution and lead to open-ended evolution in dynamic and changing environments.

Biosketch:
Prof. Dario Floreano is Director of the Laboratory of Intelligent Systems at EPFL Switzerland and Director of the Swiss National Center of Robotics. His research focuses on the convergence of biology, artificial intelligence, and robotics. He has published more than 300 peer-reviewed papers, which have been cited more than 10K times, and four books on the topics of evolutionary robotics, bio-inspired artificial intelligence, and bio-mimetic flying robots with MIT Press and Springer Verlag. He is member of the World Economic Forum Council on robotics and smart devices, co-founder of the International Society of Artificial Life, Inc. (USA), co-founder of the aerial robot company senseFly, member of the editorial board of 10 professional journals, and board member of numerous professional societies in robotics and artificial intelligence. He is also active in the public understanding of robotics and artificial intelligence, delivered more than 150 invited talks worldwide, and started the popular robotics podcast Talking Robots (now The RobotsPodcast).


GP Track Invited Speaker

Sumit Gulwani inventor of flash fill in Microsoft's Excel spreadsheet (2013) has graciously agreed to give an invited keynote for the GP track.

Sumit Gulwani
Senior Researcher
Microsoft Research
research.microsoft.com
/en-us/um/people/sumitg/

Applications of Program Synthesis to End-User Programming and Intelligent Tutoring Systems

Abstract:
Computing devices have become widely available to billions of end users, yet a handful of experts have the needed expertise to program these devices. Automated program synthesis has the potential to revolutionize this landscape, when targeted for the right set of problems and when allowing the right interaction model. The first part of this talk discusses techniques for programming using examples and natural language. These techniques have been applied to various end-user programming domains including data manipulation and smartphone scripting. The second part of this talk presents surprising applications of program synthesis technology to automating various repetitive tasks in Education including problem, solution, and feedback generation for various subject domains such as math and programming. These results advance the state-of-the-art in intelligent tutoring, and can play a significant role in enabling personalized and interactive education in both standard classrooms and MOOCs.

Biosketch:
Sumit Gulwani is a senior researcher in the RiSE group at Microsoft Research, Redmond, USA. He has expertise in formal methods and automated program analysis and synthesis techniques. As part of his vision to empower masses, he has recently focused on cross-disciplinary areas of automating end-user programming (for various systems like spreadsheets, smartphones, and robots), and building intelligent tutoring systems (for various subject domains including programming, logic, and math). Sumit's programming-by-example work led to the famous Flash Fill feature of Microsoft Excel 2013 that is used by hundreds of millions of people. Sumit obtained his PhD in Computer Science from UC-Berkeley in 2005, and was awarded the ACM SIGPLAN Outstanding Doctoral Dissertation Award. He obtained his BTech in Computer Science and Engineering from IIT Kanpur in 2000, and was awarded the President's Gold Medal.

 

 
Deep Learning and Cultural Evolution
Bridging Natural and Artificial Evolution
Applications of Program Synthesis to End-User Programming and Intelligent Tutoring Systems
 
 
 
GECCO 2013 site   GECCO 2012 site      GECCO 2011 site    GECCO 2010 site    GECCO 2009 site   GECCO 2008 site   
    GECCO 2007 site     GECCO 2006 site    GECCO 2005 site        GECCO 2004 site    GECCO 2003 site  
     GECCO 2002 site      GECCO 2001 site      GECCO 2000 site      GECCO 1999 site