Sample-Efficient Real-World Dexterous Policy Fine-Tuning via Action-Chunked Critics and Normalizing Flows
Chenyu Yang, Denis Tarasov, Davide Liconti, Hehui Zheng, Robert K. Katzschmann
01
Publications
Research across offline reinforcement learning, in-context RL, robotics, and adjacent applications.
NinA: Normalizing Flows in Action. Training VLA Models with Normalizing Flows
Denis Tarasov, Alexander Nikulin, Ilya Zisman, Albina Klepach, Nikita Lyubaykin, Andrei Polubarov, Alexander Derevyagin, Vladislav Kurenkov
cadrille: Multi-modal CAD Reconstruction with Online Reinforcement Learning
Maksim Kolodiazhnyi, Denis Tarasov, Dmitrii Zhemchuzhnikov, Alexander Nikulin, Ilya Zisman, Anna Vorontsova, Anton Konushin, Vladislav Kurenkov, Danila Rukhovich
Latent action learning requires supervision in the presence of distractors
Alexander Nikulin, Ilya Zisman, Denis Tarasov, Nikita Lyubaykin, Andrei Polubarov, Igor Kiselev, Vladislav Kurenkov
Vintix: Action Model via In-Context Reinforcement Learning
Andrey Polubarov, Nikita Lyubaykin, Alexander Derevyagin, Ilya Zisman, Denis Tarasov, Alexander Nikulin, Vladislav Kurenkov
Yes, Q-learning helps offline in-context RL
Denis Tarasov, Alexander Nikulin, Ilya Zisman, Albina Klepach, Andrei Polubarov, Nikita Lyubaykin, Alexander Derevyagin, Igor Kiselev, Vladislav Kurenkov
Object-Centric Latent Action Learning
Albina Klepach, Alexander Nikulin, Ilya Zisman, Denis Tarasov, Alexander Derevyagin, Andrei Polubarov, Nikita Lyubaykin, Vladislav Kurenkov
N-gram induction heads for in-context RL: Improving stability and reducing data needs
Ilya Zisman, Alexander Nikulin, Viacheslav Sinii, Denis Tarasov, Nikita Lyubaykin, Andrei Polubarov, Igor Kiselev, Vladislav Kurenkov
The Role of Deep Learning Regularizations on Actors in Offline RL
Denis Tarasov, Anja Surina, Caglar Gulcehre
Is Value Functions Estimation with Classification Plug-and-play for Offline Reinforcement Learning?
Denis Tarasov, Kirill Brilliantov, Dmitrii Kharlapenko
Distilling LLMs' Decomposition Abilities into Compact Language Models
Denis Tarasov, Kumar Shridhar
Offline RL for generative design of protein binders
Denis Tarasov, Ulrich Mbou Sob, Miguel Arbesu, Nima H. Siboni, Sebastien Boyer, Andries Smit, Oliver Bent, Arnu Pretorius, Marcin Skwark
Revisiting the Minimalist Approach to Offline Reinforcement Learning
Denis Tarasov, Vladislav Kurenkov, Alexander Nikulin, Sergey Kolesnikov
CORL: Research-oriented Deep Offline Reinforcement Learning Library
Denis Tarasov, Alexander Nikulin, Dmitry Akimov, Vladislav Kurenkov, Sergey Kolesnikov
Katakomba: Tools and Benchmarks for Data-Driven NetHack
Vladislav Kurenkov, Alexander Nikulin, Denis Tarasov, Sergey Kolesnikov
Anti-Exploration by Random Network Distillation
Alexander Nikulin, Vladislav Kurenkov, Denis Tarasov, Sergey Kolesnikov
Q-Ensemble for Offline RL: Don't Scale the Ensemble, Scale the Batch Size
Alexander Nikulin, Vladislav Kurenkov, Denis Tarasov, Dmitriy Akimov, Sergey Kolesnikov
Let Offline RL Flow: Training Conservative Agents in the Latent Space of Normalizing Flows
Dmitriy Akimov, Vladislav Kurenkov, Alexander Nikulin, Denis Tarasov, Sergey Kolesnikov
Prompts and Pre-Trained Language Models for Offline Reinforcement Learning
Denis Tarasov, Vladislav Kurenkov, Sergey Kolesnikov
Fixing 1-bit Adam and 1-bit LAMB algorithms
Denis Tarasov, Vasily Ershov
Predicting ethnicity with data on personal names in Russia
Alexey Bessudnov, Denis Tarasov, Viacheslav Panasovets, Veronica Kostenko, Ivan Smirnov, Vladimir Uspenskiy
02
Experience
-
Researcher, ETH Zurich Soft Robotics Lab
Research in RL for robotics. -
Research Assistant, EPFL CLAIRe Lab
Research in RL exploration with VLM models and offline RL. Supervised by Caglar Gulcehre. -
Research Assistant, ETH Zurich Language, Reasoning and Education Lab
Research in reasoning and RLAIF techniques. Supervised by Mrinmaya Sachan. -
Research Intern, InstaDeep
Research on applying offline RL techniques to drug design. -
Research Scientist, Tinkoff AI
Offline RL research, publications at ICML/NeurIPS, and development of the CORL library. -
Research Engineering Intern, Meta
AI Applied Research Relevance team; improved entity-linking model performance with distillation and unlabeled signals. -
Research Engineering Intern, Yandex
Research on distributed training over slow networks; fixes for 1-bit Adam/LAMB divergence issues. -
Research Intern, JetBrains Research
Deep learning for antibodies CDR-H3 structure prediction with improved physical validity.
03
Other
Education
- PhD in AI and Robotics, ETH Zurich (2026 - present)
- MSc in Artificial Intelligence, ETH Zurich, GPA 5.5/6 (2023 - 2025)
- BSc in Applied Math and CS, Constructor University Bremen (2022 - 2023)
- BSc in Applied Math and CS, HSE, GPA 8.73/10 (2019 - 2022)
Awards
- Best Student Paper Award, AI4D3 NeurIPS Workshop (2023)
- Winner, Ya Professional Olympiad AI Track (2022)
- 2nd place, Huawei Metric Learning Hackathon at RAAI Summer School (2019)
Service
- Reviewer: ICML, ICLR, AAAI, NeurIPS (2023-2025)
- Bachelor's diploma project supervision, Constructor University Bremen (2023-2024)