Daniel Furelos-Blanco

Postdoc @ Imperial College London.

I am a recent Ph.D. graduate in Computing from the SPIKE Research Group at Imperial College London, where I was advised by Alessandra Russo, Krysia Broda, Mark Law and Anders Jonsson. My research focused on the learning and exploitation of (hierarchies of) finite-state machines in reinforcement learning, using them as a means for task decomposition and temporal abstraction.

During my Ph.D. studies I interned at InstaDeep, where I worked on solving combinatorial optimization problems using reinforcement learning. Before starting my Ph.D., I was a research assistant in the Artificial Intelligence and Machine Learning Group at Universitat Pompeu Fabra, where I worked on enabling action concurrency in multiagent planning and temporal planning.

Selected Publications

  1. ICML
    Hierarchies of Reward Machines
    Daniel Furelos-Blanco, Mark Law, Anders Jonsson, Krysia Broda, and Alessandra Russo
    International Conference on Machine Learning (ICML), 2023
  2. NeurIPS
    Winner Takes It All: Training Performant RL Populations for Combinatorial Optimization
    Nathan Grinsztajn, Daniel Furelos-Blanco, Shikha Surana, Clément Bonnet, and Thomas D. Barrett
    Neural Information Processing Systems (NeurIPS), 2023
  3. JAIR
    Induction and Exploitation of Subgoal Automata for Reinforcement Learning
    Daniel Furelos-Blanco, Mark Law, Anders Jonsson, Krysia Broda, and Alessandra Russo
    Journal of Artificial Intelligence Research (JAIR), 2021
  4. AAAI
    Solving Multiagent Planning Problems with Concurrent Conditional Effects
    Daniel Furelos-Blanco, and Anders Jonsson
    AAAI Conference on Artificial Intelligence (AAAI), 2019