QLS

Europe/Rome 2020-06-09 15:00:00 2020-06-09 16:00:00 Statistical Learning Theory of geometrically structured data Join Zoom Meeting https://zoom.us/j/475819702 Meeting ID: 475-819-702 To obtain the PASSWORD for this zoom meeting, please contact qls@ictp.it   The traditional approach of statistical physics to supervised learning routinely assumes unrealistic generative models for the data: usually inputs are independent random variables, uncorrelated with their labels. Only recently, statistical physicists started to explore more complex forms of data, such as equally-labelled points lying on (possibly low dimensional) object manifolds. Here we provide a bridge between this recently-established research area and the framework of statistical learning theory, a branch of mathematics devoted to inference in machine learning. The overarching motivation is the inadequacy of the classic rigorous results in explaining the remarkable generalization properties of deep learning. We propose a way to integrate physical models of data into statistical learning theory, and address, with both combinatorial and statistical mechanics methods, the computation of the Vapnik-Chervonenkis entropy, which counts the number of different binary classifications compatible with the loss class. As a proof of concept, we focus on kernel machines and on two simple realizations of data structure introduced in recent physics literature: k-dimensional simplexes with prescribed geometric relations and spherical manifolds (equivalent to margin classification). Entropy, contrary to what happens for unstructured data, is nonmonotonic in the sample size, in contrast with the rigorous bounds. Moreover, data structure induces a novel transition beyond the storage capacity, which we advocate as a proxy of the nonmonotonicity, and ultimately a cue of low generalization error. The identification of a synaptic volume vanishing at the transition allows a quantification of the impact of data structure within replica theory, applicable in cases where combinatorial methods are not available, as we demonstrate for margin learning.   Zoom Meeting ICTP pio@ictp.it 9 Jun 2020

Europe/Rome 2020-11-02 07:00:00 2020-11-13 21:00:00 School on Collective Animal Behaviour *POSTPONED* | (smr 3432) Note: ICTP has suspended all educational activities from 24 February to 15 May as precautionary measures due to COVID-19; more details at https://www.ictp.it/about-ictp/media-centre/news/2020/2/ictp_suspend_activities.aspx   The School has been POSTPONED TO 2 - 13 NOVEMBER 2020     <span style="font-size:11.0pt;font-family:"Times New Roman",serif; color:red;mso-ansi-language:EN-IE">  Collective behavior is a distinctive trait of human and animal societies. The School aims at providing a thorough introduction to basic notions and advanced topics in this subject from the viewpoints of physics, biology, computer science and robotics.   Provisional list of Speakers: Iain Couzin, Max Planck Institute of Animal Behavior, Konstanz Deborah Gordon, Stanford University Marcelo Magnasco, Rockefeller University, New York Melanie Moses, University of New Mexico David Sumpter, University of Uppsala Guy Theraulaz, CRCA, Toulouse     Havana - Cuba ICTP pio@ictp.it 2 Nov 2020 - 13 Nov 2020

Europe/Rome 2020-11-30 07:00:00 2020-12-11 21:00:00 Winter School on Quantitative Systems Biology: Quantitative Approaches in Ecosystem Ecology | (smr 3486) ICTP ICTP pio@ictp.it 30 Nov 2020 - 11 Dec 2020