Michal Valko : Talks

Upcoming talks

September 21, 2026
LIX Colloquium LIX, École Polytechnique, Palaiseau, France
September 23-24, 2026
AI Horizons 2026 CUBEX, Prague, Czechia
September 24, 2026
Future Week 26 When AI Connects the Dots: The Future Beyond LLMs Atlantis, Media City Bergen, Bergen, Norway
September 24, 2026
AI ON TAP - Creandum x BOOOM Oktoberfest Schützen-Festzelt, Munich, Germany
September 25, 2026
European Researchers' Night: The Power of Science - How We Can Shape Our Future Slovakia

2027

2026

2025

2024

2023

2022

2021

  • Bootstrap your own latent - DataFest Yerevan, September 11, 2021 (DataFest 2021)
  • Bootstrapped representation learning on graphs - Turkey Machine Learning Summer School, June, 23 2021, Science Academy of Turkey (BAYÖYO 2021)
    Video: Bootstrapped representation learning on graphs
  • Self-supervised learning, BYOL, and friends - DeepMind Theory meets Practice, 2021, London, UK (DeepMind TmP 2021)
  • Graphs in Machine Learning - Medical University Graz, Austria, June 2021 (HCAI 2021)
  • Bootstrap Your Own Latent: A new approach to self-supervised learning - MIST conference in Rajecká Lesná, January, 2021 (MIST 2021)

2020

2019

  • Graphs are the new gold: The power of graphs in speeding up online learning and decision making - Global Innovation Forum, October 16-18, 2019, Yerevan, FAST, Armenia (GIF 2019)
  • Gaussian process optimization with adaptive sketching: Scalable and no regret - Recent developments in kernel methods, September 26-27, 2019, UCL, London, UK (Gatsby 2019)
  • Rotting bandits are not harder than stochastic ones - Lancaster and DeepMind Bandit Workshop, September 25-26, 2019, London, UK (LanDeep 2019)
  • Graphs are the new gold: The power of graphs in speeding up online learning and decision making - Cisco, July 23, 2019, Kraków, Poland (Cisco 2019)
  • How the negative dependence broke the squadratic barrier for learning with graphs and kernels - Yandex HQ, July 5, 2019, Moscow, Russia (Yandex 2019)
  • Active multiple matrix completion with adaptive confidence sets - RAAI Summer School 2019, July 3-8, 2019, Moscow Institute of Physics and Technology (RAAI 2019)
    Video: Active multiple matrix completion with adaptive confidence sets
  • Maximum likelihood estimation for low rank DPPs - ICML workshop on negative dependence, June 14-15, 2019, Long Beach, USA (ICML 2019 DPPs)
  • How the negative dependence broke quadratic barrier for learning with graphs and kernels. - ICML workshop on negative dependence, June 14-15, 2019, Long Beach, USA (ICML 2019)
  • Optimal bandit submodular minimization - ICML 2019 workshop, June 9-15, 2019, Long Beach, USA (ICML 2019 submod)
  • Active block-matrix completion with adaptive confidence sets - Theoretical Computer Science seminar, May 28, 2019, Comenius University in Bratislava, Slovakia (CU May 2019)
  • 10 year road to breaking the quadratic barrier for graphs and matrices - Theoretical Computer Science seminar, February 22, 2019, Comenius University in Bratislava, Slovakia (CU Feb 2019)
  • Graphs are the new gold - Data Analytics Meetings, February 20, 2019, P.J. Šafárik University in Košice, Slovakia (UPJS 2019)
  • The power of graphs in speeding up online learning and decision making - Department of Pure Mathematics and Mathematical Statistics, January 25, 2019, University of Cambridge, UK (Cambridge 2019)
  • The power of graphs in speeding up online learning and decision making - Verimag, January 8, 2019, CNRS Grenoble, France (CNRS Jan8 2019)
  • A simple parameter-free and adaptive approach to optimization under a minimal local smoothness assumption - Verimag, January 7, 2019, CNRS Grenoble, France (CNRS Jan7 2019)

2018

  • The power of graphs in speeding up online learning and decision making - DeepMind, October 23, 2018, London, UK (DeepMind 2018)
  • Active block-matrix completion with adaptive confidence sets - International Workshop Optimization and Machine Learning, September 10-13, 2018, CIMI, Toulouse (CIMI 2018)
  • Online influence maximization - Workshop on Graph Learning, May 14, 2018, LINCS, Paris (LINCS 2018)
  • Recommender systems - Big Data Day, Polytech'Lille, March 22, 2018 (Polytech'Lille 2018)
  • Pliable rejection sampling - GDR Isis, Télécom ParisTech in Paris, February 8, 2018 (ISIS 2018)
  • Graph Bandits - MIST conference in Rajecká Lesná, January 7, 2018 (MIST 2018)

2017

  • SequeL, graphs in ML, and online recommender systems - R&DV du Plateau Inria 2017, November 9, 2017, Lille, France (Euratechnologies 2017)
  • Sequential sampling for kernel matrix approximation and online learning - DeepMind, September 19, 2017, London, UK (DeepMind 2017)
  • Active learning on networks and online influence maximization - Decision Theory and Network Science: Methods and Applications, September 18, 2017, Lancaster, UK (STOR-i 2017)
  • Side observation in graph bandits - ICML 2017 workshop on Picky Learners, August 11, 2017, Sydney, Australia (ICML 2017)
  • Efficient second-order online kernel learning with adaptive embedding (Impromptu talk) - Conference on Learning Theory, July 9, 2017, Amsterdam, Netherlands (COLT 2017)
  • Distributed sequential sampling for kernel matrix approximation - Toulouse Institute of Mathematics, France, June 29, 2017 (IMT 2017)
  • Where Is Justin Bieber? Or Online Influence Maximization - Inria Scientific Days 2017, June 14, 2017, Sophia-Antipolis, France (JS 2017)
  • Where Is Justin Bieber Hiding? Or How to Maximize Influencer Detection on Social Networks? (Popularization invited talk) - Inria Lille, France, May 30, 2017 (13:45 2017)
  • Dating day 2017, March 30, 2017, Lille, France (Dating 2017)
  • Distributed sequential sampling for kernel matrix approximation - Joint seminar of MuSyAD group of Universität Potsdam and Amazon at Amazon Berlin, Germany, March 22, 2017 (Berlin 2017)

2016

Demos and Presentations

  • Michal Valko, Miloš Hauskrecht, G. Cooper, S. Visweswaran, M. Saul, A. Seybert, J. Harrison, A. Post: Conditional Anomaly Detection - CS Day 2008 (CS Day 2008) [["#1st by people"], ["#2nd by faculty"]] also at (University of Pittsburgh, Arts & Sciences (Grad Expo 2008))
  • Michal Valko: Conditional anomaly detection with adaptive similarity metric - CS Department Research Competition (Research 2008) [#1st place]
  • Michal Valko, Branislav Kveton, Matthai Philiposse: Robust Face Recognition Using Online Learning - 9th University of Pittsburgh Science conference (SCIENCE 2009) also at (Live Demonstration (Grad Expo 2010) and (CS Day 2010))
  • Branislav Kveton, Michal Valko, Matthai Philiposse: Real-Time Adaptive Face Recognition - 23rd Neural Information Processing Systems conference (NeurIPS 2009)
  • Michal Valko: Graph-Based Anomaly Detection with Soft Harmonic Functions - CS Department Research Competition (Research 2011) [#1st place] also at (Grad Expo 2011 and CS Day 2011)

2015 and before

  • Online decision-making on graphs - DaSciM, LIX, École Polytechnique, France, April 14, 2015 (X 2015)
  • Bandits on Graphs: Exploiting Smoothness and Side Observations - CMLA at ENS Cachan, France, December 16, 2014 (ENS 2014)
  • Optimistic Optimization - MIST conference in Fačkovské sedlo, January 7, 2014 (MIST 2014)
  • Sequential Face Recognition with Minimal Feedback - 30 minutes of Science, Lille, May 2, 2013 (30MIN 2013)
  • Stochastic Simultaneous Optimistic Optimization - ComplACS Workshop, Leoben, Austria, November 2012 (ComplACS 2012)
  • One Class Learning From Streams of Unlabeled Data - Large-scale Online Learning and Decision Making, September 17, 2012 (LSOLDM 2012)
  • Scaling Graph-Based Algorithms - LAMPADA workshop, July 20, 2012 (LAMPADA 2012)
  • Large Scale Sequential Learning - Slovak Oxford Science, April 8, 2012 (Oxford UK 2012)
  • Adaptive Graph-Based Algorithms - Microsoft Research Redmond, July 6, 2011 (MSR 2011)
  • Online Semi-Supervised Learning - MPI Tuebingen, Germany, July 2011 (MPI 2011)
  • Semi-supervised Learning with Random Walks on Graphs - 6th Comenius University Alumni conference (TAM 2009)
  • Stochastic Simultaneous Optimistic Optimization - International Conference on Machine Learning (ICML 2013), June 2013, Atlanta, Georgia, USA (ICML StoSOO 2013)
  • Conditional Anomaly Detection Methods for Analysis of Clinical Alert Data - Machine Learning in Health Care Applications, ICML 2008, July 9, 2008, Helsinki, Finland (ICML Health 2008)
  • Distance Metric Learning for Conditional Anomaly Detection - FLAIRS, May 2008, Coconut Grove, Florida, USA (FLAIRS 2008)
  • Learning Predictive Models for Combinations of Heterogeneous Proteomic Data Sources - AMIA STB, March 2008, San Francisco, California, USA (AMIA STB 2008)
  • Dimensionality Reduction of High-throughput Proteomic Data using Partial Discriminative Projections - AI Seminar, FMFI, May 23, 2006 (FMFI AI Seminar 2006)