Data Science Seminars

Shape Constraints Meet Kernel Machines

Speaker: Zoltan Szabo (London School of Economics)

Shape constraints (such as non-negativity, monotonicity, convexity, or supermodularity) provide a principled way to encode prior information in predictive models with numerous successful applications in econometrics, finance, biology, reinforcement learning, and game theory. Incorporating this side information in a hard way (for instance at all point of an interval) however is an extremely challenging problem. In this talk I am going to present a unified and modular convex optimization framework to encode hard affine constraints on function values and derivatives into the flexible class of reproducing kernel Hilbert spaces. The efficiency of the technique is illustrated in the context of joint quantile regression (analysis of aircraft departures), convoy localization and safety-critical control (piloting an underwater vehicle while avoiding obstacles). [This is joint work with Pierre-Cyril Aubin-Frankowski.]

Time and Place: Thursday 04th November at 3 pm (online)

Previous talks

Toward a Perpetual Learning Machine in Continual Control

Speaker: Shane Gu (Google Brain)


Monotonic Alpha-Divergence Variational Inference

Speaker: Kamélia Daudel (University of Oxford)


Modeling Knowledge Incorporation into Topic Models and their Evaluation

Speaker: Silvia Terragni (University of Milano-Bicocca)


Explainable Fact Checking for Statistical and Property Claims

Speaker: Paolo Papotti (Professor at EURECOM)


Interpretable Comparison of Generative Models

Speaker: Wittawat Jitkrittum (Research Scientist at Google Research)


Explaining the Explainer: A First Theoretical Analysis of LIME

Speaker: Damien Garreau (Assistant Professor at the University Cote d’Azur)


Variable Prioritization in Nonlinear Black Box Methods, with application in Genomics and to Interpreting Deep Neural Network

Speaker: Seth Flaxmann (Lecturer in the Statistics at Department of Mathematics of Imperial College London)


Learning on Aggregate Outputs with Kernels

Speaker: Dino Sejdinovic (Associate Professor at the Department of Statistics, University of Oxford)


Sparse Approximate Inference for Spatio-Temporal Point Process Models with Application to Armed Conflict

Speaker: Andrew Zammit Mangion (Senior Research Fellow at the University of Wollongong - NIASRA, Australia)