Dr. Michal Valko
Facts
35 properties- Instance of
- human
- Place of birth
- Košice
- Occupation
- university teacher, computer scientist, researcher, artificial intelligence researcher, entrepreneur, angel investor[10]
- Employer
- Google DeepMind (2019–2024), Institut National de Recherche en Informatique et en Automatique (2011–present), Meta Platforms (2024–2024), University of Pittsburgh (2005–2011), Intel (2009–2010), Slovak Academy of Sciences (2003–2006), École normale supérieure Paris-Saclay (2014–present), Isara Labs (2025–present)[28][29][30]
- Field of work
- machine learning, artificial intelligence, deep learning, game theory, multi-armed bandit, natural language processing, self-supervised learning, large language model, statistics, computer science, reinforcement learning from human feedback, preference learning[11][12][13][14][15][16][17][18][19][20][21][22][23][24]
+6 more
Monte Carlo tree search, graph theory, recommender system, online machine learning, kernel method, mathematical optimization - Position held
- Principal Llama Scientist (2024–2024), Founding Researcher (2025–present), External Lecturer (2014–present), ESET Science Award Envoy (2021)[25][26][27]
- Educated at
- University of Pittsburgh (2005–2011), NOVA University Lisbon (2005–2005), Comenius University (2000–2005), Gymnasium Alejová (1992–2000), École normale supérieure Paris-Saclay (2015–2016), Carnegie Mellon University (2007–2009)
- Award received
- Inria Prime d'excellence scientifique (2014, 2018), ICML Top 10 Reviewer Award (2018), ICML Reviewer Award (2015), Distinguished Alumni of Comenius University (2015), Google Best Paper Award (2010), Homer Warner Award (2010), Compunetix Best Research Award (2008, 2011), University of Pittsburgh Honors Convocation Recognition (2009), Andrew Mellon Predoctoral Fellowship (2008), Forbes Changemakers (2024), ICML Outstanding Paper Award (2023), ASAI AI Social Impact of the Year (2026)[32][33][34][35][36][37][38][39][40][41][42][43]
- Notable work
- Bootstrap your own latent: A new approach to self-supervised Learning, Nash Learning from Human Feedback, BYOL-Explore: Exploration by Bootstrapped Prediction, Large-Scale Representation Learning on Graphs via Bootstrapping, Bandits on Graphs and Structures, Llama 3, The Miscible-Immiscible Quantum Phase Transition in Coupled Two-Component Bose-Einstein Condensates in 1D Optical Lattices, Finite-Time Analysis of Kernelised Contextual Bandits, Gamification of Pure Exploration for Linear Bandits, The AI Book: Prepare for the future, Adapting to game trees in zero-sum imperfect information games, Human alignment of large language models through online preference optimisation
+146 more
Generalized preference optimization: A unified approach to offline alignment, Fast rates for maximum entropy exploration, Quantile credit assignment, UCB Momentum Q-learning: Correcting the bias without forgetting, From Dirichlet to Rubin: Optimistic exploration in RL without bonuses, Optimistic Posterior Sampling for Reinforcement Learning: Worst-Case Regret Bounds, Model-free learning for two-player zero-sum partially observable Markov games, Stochastic shortest path: minimax, parameter-free and towards horizon-free regret, The Llama 3 Herd of Models, A general theoretical paradigm to understand learning from human preferences, Broaden Your Views for Self-Supervised Video Learning, Retrieval-augmented reinforcement learning, Self-Supervised Representation Learning Using Bootstrapped Latent Representations, Reinforcement Learning by Directly Learning an Advantage Function, Reinforcement Learning Using Hindsight to Model Unpredictable Aspects of the Future, Planning in entropy-regularized Markov decision processes and games, Monte-Carlo tree search as regularized policy optimization, Imp: Highly Capable Large Multimodal Models for Mobile Devices, DPPy: Sampling determinantal point processes with Python, Gaussian process optimization with adaptive sketching, Spectral Thompson Sampling, Etale covers of Kawamata log terminal spaces and their smooth loci, BYOL works even without batch statistics, Local and Adaptive Mirror Descents in Extensive-Form Games, Decoding-time realignment of language models, VA-learning as a more efficient alternative to Q-learning, Curiosity in hindsight: Intrinsic exploration in stochastic environments, Understanding self-predictive learning for reinforcement learning, Demonstration-regularized RL, Unlocking the power of representations in long-term novelty-based exploration, Accelerating Nash learning from human feedback via Mirror Prox, Extropy: Complementary Dual of Entropy, Mine Your Own vieW: Self-supervised learning through across-sample prediction, Maximum Entropy Semi-Supervised Inverse Reinforcement Learning, Conditional Anomaly Detection with Soft Harmonic Functions, Online Semi-Supervised Learning on Quantized Graphs, Middle-mile logistics through the lens of goal-conditioned reinforcement learning, Revealing graph bandits for maximizing local influence, Influence maximization of informed agents in social networks, Optimal design for reward modeling in RLHF, Understanding the performance gap between online and offline alignment, Preference optimization with multi-sample comparisons, RL-finetuning LLMs from on- and off-policy data with a single algorithm, A new bound on the cumulant generating function of Dirichlet processes, Sharp deviations bounds for Dirichlet weighted sums with application to analysis of Bayesian algorithms, KL-entropy-regularized RL with a generative model is minimax optimal, On the approximation relationship between optimizing ratio of submodular (RS) and difference of submodular (DS) functions, Learning to Act Greedily: Polymatroid Semi-Bandits, The Harder Path: Last iterate convergence for uncoupled Learning in zero-sum games with bandit feedback, rlberry: A Reinforcement Learning Library for Research and Education, Model-free posterior sampling via learning rate randomization, DoMo-AC: Doubly multi-step off-policy actor-critic algorithm, Regularization and variance-weighted regression achieves minimax optimality in linear MDPs: Theory and practice, Half-Hop: A graph upsampling approach for slowing down message passing, Scaling Gaussian process optimization by evaluating a few unique candidates multiple times, Adaptive multi-goal exploration, Marginalized operators for off-policy reinforcement learning, Drop, Swap, and Generate: A Self-Supervised Approach for Generating Neural Activity, A provably efficient sample collection strategy for reinforcement learning, Unifying gradient estimators for meta-reinforcement learning via off-policy evaluation, Fast active learning for pure exploration in reinforcement learning, Revisiting Peng's Q(λ) for modern reinforcement learning, Taylor expansion of discount factors, Online A-optimal design and active linear regression, Kernel-based reinforcement Learning: A finite-time analysis, Game Plan: What AI can do for Football, and What Football can do for AI, A kernel-based approach to non-stationary reinforcement learning in metric spaces, Episodic reinforcement learning in finite MDPs: Minimax lower bounds revisited, Sample complexity bounds for stochastic shortest path with a generative model, Adaptive reward-free exploration, Fast sampling from $$\beta $$-ensembles, Density-based bonuses on learned representations for reward-free exploration in deep reinforcement learning, Geometric entropic exploration, Improved sample complexity for incremental autonomous exploration in MDPs, Sampling from a k-DPP without looking at all items, Statistical efficiency of Thompson sampling for combinatorial semi-bandits, Planning in Markov decision processes with gap-dependent sample complexity, Taylor expansion policy optimization, Budgeted online influence maximization, No-regret exploration in goal-oriented reinforcement learning, Improved sleeping bandits with stochastic action sets and adversarial rewards, Stochastic bandits with arm-dependent delays, Near-linear time Gaussian process optimization with adaptive batching and resparsification, Covariance-adapting algorithm for semi-bandits with application to sparse rewards, Fixed-confidence guarantees for Bayesian best-arm identification, Adaptive multi-fidelity optimization with fast learning rates, Derivative-free & order-robust optimisation, A single algorithm for both restless and rested rotting bandits, Reward-free exploration beyond finite-horizon, Spectral Bandits, Multiagent evaluation under incomplete information, Exact sampling of determinantal point processes with sublinear time preprocessing, On two ways to use determinantal point processes for Monte Carlo integration, Scale-free adaptive planning for deterministic dynamics & discounted rewards, Exploiting structure of uncertainty for efficient matroid semi-bandits, A simple dynamic bandit-based algorithm for hyper-parameter tuning, Rotting bandits are not harder than stochastic ones, Active multiple matrix completion with adaptive confidence sets, Finding the bandit in a graph: Sequential search-and-stop, A simple parameter-free and adaptive approach to optimization under a minimal local smoothness assumption, General parallel optimization without metric, Les processus ponctuels déterminantaux en apprentissage automatique, Optimizing human learning workshop eliciting adaptive sequences for learning (WeASeL), Optimistic optimization of a Brownian, Adaptive black-box optimization got easier: HCT needs only local smoothness, Compressing the Input for CNNs with the First-Order Scattering Transform, Improved large-scale graph learning through ridge spectral sparsification, Best of both worlds: Stochastic & adversarial best-arm identification, Efficient second-order online kernel learning with adaptive embedding, Online Influence Maximization under Independent Cascade Model with Semi-Bandit Feedback, Second-order kernel online convex optimization with adaptive sketching, Zonotope hit-and-run for efficient sampling from projection DPPs, Distributed adaptive sampling for kernel matrix approximation, Trading off rewards and errors in multi-armed bandits, Blazing the trails before beating the path: Sample-efficient Monte-Carlo planning, Pack only the essentials: Adaptive dictionary learning for kernel ridge regression, Rewards and Errors in Multi-armed Bandit for Interactive Education, Pliable rejection sampling, Analysis of Nyström method with sequential ridge leverage scores, Online learning with Erdős-Rényi side-observation graphs, Online learning with noisy side observations, Black-box optimization of noisy functions with unknown smoothness, Simple regret for infinitely many armed bandits, Cheap Bandits, Large-scale semi-supervised learning with online spectral graph sparsification, Efficient Learning by Implicit Exploration in Bandit Problems with Side Observations, Extreme Bandits, Online Combinatorial Optimization with Stochastic Decision Sets and Adversarial Losses, Bandits attack function optimization, Learning from a single labeled face and a stream of unlabeled data, Outlier detection for patient monitoring and alerting, Semi-supervised apprenticeship learning, Identification of Microbial and Proteomic Biomarkers in Early Childhood Caries, Adaptive Graph-Based Algorithms for Conditional Anomaly Detection and Semi-Supervised Learning, Semi-Supervised Learning with Max-Margin Graph Cuts, Conditional outlier detection for clinical alerting, Online semi-supervised perception: Real-time learning without explicit feedback, Feature importance analysis for patient management decisions, Distance Metric Learning for Conditional Anomaly Detection, Evidence-based anomaly detection in clinical domains, Feature Selection and Dimensionality Reduction in Genomics and Proteomics, Evolutionary Feature Selection for Spiking Neural Network Pattern Classifiers, An analysis of Mirror-Descent Soft Actor-Critic, Language Generation with Replay: A Learning-Theoretic View of Model Collapse, Optimal last-iterate convergence in matrix games with bandit feedback using the log-barrier, DiG-bench: Discovery in Games - Official website
- misovalko.github.io
- Doctoral advisor
- Milos Hauskrecht[31]
- Short name
- Michal Valko
- Doctoral student
- Jean-Bastien Grill (2019), Daniele Calandriello (2017), Giorgio Racca, Côme Fiegel (2025), Daniil Tiapkin (2025), Lisa Bedin (2025), Jean Tarbouriech (2022), Omar Darwiche Domingues (2022), Xuedong Shang (2021), Julien Seznec (2021), Tomáš Kocák (2016), Guillaume Gautier, Pierre Perrault[48][49][50][51][52][53][54][55][56][57][58][59][60]
- Work location
- Paris (2011–present), San Francisco (2024–2024), Pittsburgh (2005–2011), Lille (2011–present), Lyon, London (2025–present), Bratislava (2003–2006), Lisbon (2005)
- Nominated for
- ASAI AI Innovator of the Year (2026)
- Academic degree
- habilitation to Supervise Research (2016), Doctor of Philosophy (2011), Master of Science (2005)[61][62][63]
- Academic thesis
- Adaptive Graph-Based Algorithms for Conditional Anomaly Detection and Semi-Supervised Learning, Bandits on Graphs and Structures
- Affiliation
- Google DeepMind (2019–2024), Institut National de Recherche en Informatique et en Automatique (2012–present), Doctolib, Sword Health, Entalpic AI, Formula VC, KAYA VC, Altrove, Tortus AI, Forgent AI, Robi Labs
- Birth name
- Michal Valko
- Copyright status as a creator
- works protected by copyrights
- Described by source
- Denník N, The Wall Street Journal, Business Insider, New York Post, Forbes, Radio and Television of Slovakia, Aktuality.sk, United Nations Office for Digital and Emerging Technologies, Le Nouvel Obs, SME, .týždeň, Hospodárske noviny
+27 more
TV Markíza, Euractiv, Les Échos, Sciences et Avenir, Engadget, Pravda, Refresher, Forbes Česko, The Slovak Spectator, Gulf News, Khaleej Times, Arabian Business, Rádio Expres, TA3, Rádio Devín, Zawya, France Culture, Živé.sk, The Epoch Times, PC Revue, Le Monde, Nový čas, France Inter, Radio France Internationale, BFM TV, Le Parisien, Franceinfo
Show all 35 facts +11
- Family name
- Valko
- Given name
- Michal
- Hashtag
- misovalko
- Member of
- Slovak Mathematicians and Physicists Scientific Society (2000–present), Inria Evaluation Committee (2014–2019), Association for Computing Machinery, Institut National de Recherche en Informatique et en Automatique (2014–2019), European Laboratory for Learning and Intelligent Systems[44][45][46]
- Name
- Michal Valko
- Name in native language
- Michal Valko
- Participant in
- LOD 2024 (2024), Sword AI Summit 2025 (2025), Reinforcement Learning Summer School 2019 (2019), Organization for Economic Cooperation and Development (2025), Globsec (2025), Princeton University (2024), Isaac Newton Institute (2025), Simons Institute for the Theory of Computing (2022), BratislavAI Forum 2025 (2025), AAAI Conference on Artificial Intelligence (2022), International Conference on Artificial Intelligence and Statistics (2020), International Joint Conference on Artificial Intelligence (2013)
+71 more
Ninth International Conference on Learning Representations (2021), European Workshop on Reinforcement Learning (2018), Conference on Uncertainty in Artificial Intelligence (2022), TEDx conference (2024), Artificial Intelligence Action Summit (2025), European Researchers’ Night (2025), World Artificial Intelligence Conference (2025), Runtime Verification (2022), Alan Turing Institute (2025), Flatiron Institute (2025), GHOST Day: Applied Machine Learning Conference (2025, 2026), Conference on Learning Theory (2017), International Conference on Machine Learning (2019), Eastern European Machine Learning Summer School (2023), Conference on Neural Information Processing Systems (2009), Google DeepMind (2018), Yandex (2019), Google (2020), Cisco (2019), Neural Information Processing Systems 2023 (2023), Thirty-sixth Conference on Neural Information Processing Systems (2022), Thirty-fifth Conference on Neural Information Processing Systems (2021), Thirty-fourth Conference on Neural Information Processing Systems (2020), Thirty-third Conference on Neural Information Processing Systems (2019), Thirty-second Annual Conference on Neural Information Processing Systems (2018), Thirty-first Conference on Neural Information Processing Systems (2017), Thirtieth Annual Conference on Neural Information Processing Systems (2016), Twenty-ninth Conference on Neural Information Processing Systems (2015), Twenty-eighth Conference on Neural Information Processing Systems (2014), Forty-Second International Conference on Machine Learning (2025), The Fortieth International Conference on Machine Learning (2023), The Thirty-ninth International Conference on Machine Learning (2022), The Thirty-Eighth International Conference on Machine Learning (2021), Thirty-seventh International Conference on Machine Learning (2020), Thirty-sixth International Conference on Machine Learning (2019), Thirty-fifth International Conference on Machine Learning (2018), Thirty-fourth International Conference on Machine Learning (2017), The 25th International Conference on Artificial Intelligence and Statistics (2022), The 23rd International Conference on Artificial Intelligence and Statistics (2020), European Conference on Computer Vision 2018 (2018), Twenty-Eighth AAAI Conference on Artificial Intelligence (2014), 41st International Conference on Machine Learning (2024), Thirty-eighth Annual Conference on Neural Information Processing Systems (2024), The Twelfth International Conference on Learning Representations (2024), The Web Conference 2024 (2024), Knowledge Discovery and Data Mining (2024), GLOBSEC Forum (2024), World Economic Forum Annual Meeting 2025 (2025), Paris Peace Forum (2024), Web Summit (2024), Collision (2024), VivaTech (2024), AAAI 2025 (2025), New York University Abu Dhabi (2026), Cerebras (2026), Sapienza University of Rome (2026), New Challenges in Central Banking Conference (2026) (2026), Reinforcement Learning Summer School 2026 (2026), Future Week 2026 (Bergen) (2026), AI Day 2026 (Comenius University) (2026), International Summer School on Generative AI 2026 (2026), Dynamically Evolving Long-Term Autonomy, ExTra-Learn (2018), Theoretically grounded efficient algorithms for high-dimensional and continuous reinforcement learning (2018), Beyond Online Learning for better Decision making, Bayesian statistics for expensive models and tall data, CompLACS, Educational Bandits, Non-parametric sequential prediction project, Adaptive allocation of resources for recommender systems, Sequential Machine Learning for Adaptive Educational Systems (2019) - Social media followers
- 15,000 (2024)
- Writing language
- English
- ISNI
- 0000000351150335
- ORCID iD
- 0009-0009-2673-0144
- VIAF cluster ID
- 3284171191086658030003
- OpenAlex ID
- A5106038276
- Image (Wikimedia Commons)
- Michal Valko.jpg ↗
- ACM Digital Library author ID
- 99661016893
- Crunchbase person ID
- michal-valko
- DBLP author ID
- 03/5455
- Described at URL
- www.franceinfo.fr, e.dennikn.sk/…/coho-sa-naozaj-bat-pri-ai-vyhubenie-ludstva-je-nepravdepodobne-ine-bezpecnostne-rizika-su-skutocne-hovoria-experti, vedaodpoveda.podbean.com, e.dennikn.sk/…/najvacsie-ai-laboratoria-vystrasili-svet-autonomnymi-hackerskymi-utokmi-svojich-modelov-slovensky-expert-vysvetluje-co-sa-stalo, www.itapa.sk/jarna-itapa-2026-stat-musi-chranit-deti-online-no-nesmie-brzdit-inovacie, www.linkedin.com, kinit.sk, www.tyzden.sk/…/algoritmy-uz-nevyberaju-len-obsah-coskoro-ho-budu-priamo-pre-nas-vyrabat, vedanadosah.cvtisr.sk, www.tyzden.sk/…/buducnost-firiem-nie-je-otazkou-zajtrajska-orange-konferencia-2026-ukazala-ze-sa-tvori-uz-dnes, www.youtube.com, grandmagazine.argusmedia.sk, sita.sk, 2026.rlsummerschool.com, intelligent-systems.net, www.itapa.sk/spikri/michal-valko, acdl2026.icas.events, ghostday.pl, www.slovak-oxford-science.co.uk
- Dimensions author ID
- 011674244541.25
- GitHub account
- misovalko
- Google Knowledge Graph ID
- /g/11f3_36tjr
Show all 26 identifiers +14
- Google Scholar author ID
- jrazNCQAAAAJ
- HAL author ID
- michal
- Hugging Face user or organization ID
- misovalko
- IdRef ID
- 22360934X
- Justia Patents inventor ID
- michal-valko
- LinkedIn personal profile ID
- michalvalko
- MR Author ID
- 1169452
- Mathematics Genealogy Project ID
- 156876
- Quora topic ID
- Michal-Valko
- ResearchGate profile ID
- Michal-Valko
- Scopus author ID
- 25642222500
- Semantic Scholar author ID
- 1806291
- X (Twitter) username
- misovalko (2008–present)
- ZbMATH author ID
- valko.michal
0 sources
Summary
Dr. Michal Valko was born in Košice[1]. He has worked as a university teacher[2], computer scientist[3], researcher[4], artificial intelligence researcher[5], and entrepreneur[6].
FAQs
Where was Dr. Michal Valko born?
Dr. Michal Valko's place of birth was Košice[1].
What did Dr. Michal Valko do for work?
Dr. Michal Valko worked as a university teacher[2], computer scientist[3], researcher[4], artificial intelligence researcher[5], and entrepreneur[6].
What awards did Dr. Michal Valko receive?
Honors received include Inria Prime d'excellence scientifique[32], ICML Top 10 Reviewer Award[33], and ICML Reviewer Award[34].