CogniGron Seminar: Bert Kappen (Radboud University Nijmegen) - "Future machine learning requires new computing paradigms"
|When:||Mo 09-07-2018 11:00 - 12:00|
|Where:||5159.0062 (Energy Academy Europe)|
It has been a long held dream to realize intelligence in artificial devices, such as computers, robots and cars. Recently, through improved hardware and the availability of large data, some spectacular advances have been made such as in game playing, pattern recognition and autonomous driving. It becomes clear that computing, memory and energy consumption will be the main challenges for the future growth of AI. Part of the answer is to develop ideas and devices such that the 'physics will do the work' for us. In this talk, after reviewing some recent AI succeses, I will focus on two recent ideas in this direction: stochastic neural computing and quantum machine learning. I will argue that neural networks with stochastic synapes can provide an energy efficient and feasible alternative to deep learning. Our initial research on the binary perceptron shows that this approach can yield spectacular performance in learning a NP hard problem and scales to deeper networks. Secondly, I will present a novel quantum maximum likelihood framework for learning that generalizes Bayesian inference to quantum states with a surprisingly powerful representation. As an example, I show that the quantum Boltzmann machine can learn very non-linear problems while its classical analogue cannot.
Contact & more about Bert:
Prof. Bert Kappen conducts theoretical research that lie at the interface between machine learning, control theory, statistical physics, computer science, computational biology and artifcial intelligence. He has developed many novel approximate inference methods inspired by methods from statistical physics. He has pioneered the mean field analysis of stochastic neural networks with dynamical synapses, revealing up and down states and rapid switching. He has identified a novel class of non-linear stochastic control problems that can be solved using path integrals. This approach has been adapted by leading robotics groups world wide, and is recognized as an important novel approach to stochastic control. His work on mean field theory for asymmetric stochastic neural networks is at the basis of current research to find connectivity patterns in neural circuits. He is author of about 130 peer reviewed articles in scientific journals and leading conferences. In collaboration with medical experts, he has developed a Bayesian medical expert system, including approximate inference methods, and he has co-founded the company Promedas to commercialize this system. He is director of SNN, the Dutch foundation for Neural Networks. SNN has a long reputation for successfully applying neural network and machine learning methods in collaboration with numerous industrial partners. He has co-founded the company Smart Research bv, that offers commercial service on machine learning and that has developed the Bonaparte Disaster Victim Identification software. He is honorary faculty at the Gatsby Unit for Computational Neuroscience at University College London.