Oriol Corcoll

Oriol Corcoll

Research Scientist at Spotify

University of Tartu

Biography

I am a Research Scientist at Spotify working on Causal ML. My Ph.D. from University of Tartu focused on Causal Reinforcement Learning. I studied how reinforcement learning agents can understand the world in terms of cause and effect. Building agents that see the world from a causal lens can enable principled drug discovery, informed interventions on gene regulatory networks, and better decision-making in general.

Previously @ Amazon Alexa

Interests
  • Reinforcement Learning
  • Causality
  • Actual Causality
  • Hierarchical RL
  • Learning by Explaining
Education
  • PhD in Computer Science, 2022

    University of Tartu

  • MSc in Data Science, 2018

    Queen Mary University of London

  • BSc in Computer Science, 2014

    Politechnic University of Catalonia

Publications

Quickly discover relevant content by filtering publications.
(2022). Explanatory World Models via Look Ahead Attention for Credit Assignment. Causal Representation Learning (UAI).

PDF CRL

(2022). Mind the gap: Challenges of deep learning approaches to Theory of Mind. Under review.

PDF Cite DOI Arxiv

(2022). Quantifying reinforcement-learning agent's autonomy, reliance on memory and internalisation of the environment. Entropy. Special issue Towards a Quantitative Understanding of Agency.

PDF Cite Entropy

(2022). Disentangling causal effects for hierarchical reinforcement learning. Conference on Causal Learning and Reasoning (CLeaR).

PDF Cite Arxiv

(2021). Did I do that? Blame as a means to identify controlled effects in reinforcement learning. URL @ International Conference on Machine Learning (ICML) and Transactions of Machine Learning Research (TMLR).

PDF Cite Poster Arxiv

Contact