Abstract
Neural networks provide new tools, new questions and new foundations for solving practical problems in prediction, modeling, decision and control, state estimation, pattern recognition, data mining, etc. Traditional statistics tries to collect a huge library of different methods for different tasks, but the brain is living proof that one system can do it all, if there is data. It proves that a system can manage many millions of variables without being confused (usually, when we really use our brain). Neural-type learning implemented on distributed elements like neurons allows use of new general chips like CNN, with many thousand times more throughput than older chip types. Adaptive dynamic programming (ADP) is a type of learning related to the Bellman equation, which is as important to the general adaptive optimal value-based management of complex systems (like energy systems) as Maxwell's Laws and the Dirac equation are to electronics. New advances, designs and applications of ADP on neural networks show some big improvements in performance and stability over older methods, and a new functional way to understand the brain and mind themselves. Still, this new hyper-rational way to understand the mind is very synergistic with practical ways to understand the mind rooted in personal experience which have evolved in wise folk traditions like Freud, Jung, Taoism, Confucianism and even some types of Buddhism.
Speaker
Paul J. Werbos holds four degrees from Harvard and the London School of Economics in: (1) economics; (2) international political systems, emphasizing European economic institutions; (3) applied mathematics, with a major in quantum physics and a minor in decision and control; (4) applied mathematics, towards an interdisciplinary Ph.D. thesis. Prior to that, during high school, he obtained an FCC First Class Commercial Radiotelephone License, and took undergraduate and graduate mathematics courses at Princeton and the University of Pennsylvania.
For about four years after the PhD, he taught courses at Maryland in quantitative methods and global futures, and performed research in intelligent systems for policy application. Then for nine years he worked at the Department of Energy evaluating and developing a wide range of energy forecasting models. In 1989 he joined NSF as a program director in the ECS Division managing Neuroengineering (neural networks) and Emerging Technology Initiation. Within CNCI area, his main goal is to maximize the development and dissemination of step-by-step advances in systems design which will lead to an understanding and replication of the general kind of learning-based intelligence described in the recent NSF workshop on learning and approximate dynamic programming. In essence, this involves an equal emphasis on general-purpose intelligent control issues, and on the prediction and learning issues important to developing the subsystems needed for such an integrated intelligent system. He also has a special interest in the use of computational intelligence to help provide new options in large-scale important areas such as energy, sustainability, and the like. He has also initiated a new area of quantum neural networks or quantum learning, an approach to the algorithm gap in quantum computing.
He has served as President of the International Neural Network Society, where he is still on the Governing Board. He also serves on the AdCom of the IEEE Industrial Electronics Society, and the chairs the Awards Committee of the IEEE Neural Network Society. He is also on the Planning Committee of the Millennium Project of the United Nations University (http://millennium-project.org), and on various cross-cutting working groups such as the working group on energy production and distribution of the interagency Climate Change Technology Program.