Reinforcement Learning

  • M. Conserva
    Total War Simulator [Code]    Star Fork
    The Total War Simulator is a research project that aims to reproduce battles from the Total War game series for reinforcement learning agents training. The latest version is implemented in Unity with an earlier version available for Python. The Total War simulator includes melee and ranged units, human control and a Python AI interface with the Unity MLAgents toolkit.

  • Alberto Soragna and M. Conserva
    Deep Briscola [Code]    Star Fork
    Deep Briscola is a research project that explores the effectiveness of state-of-the-art reinforcement learning agents to play the Italian card game of Briscola.

  • M. Conserva, Alessia Eletti, Pauline de Lavallade, and Jaeik Jeon
    Learn to Read to Read to Learn [Report] [Code]
    Group project for Natural Language Processing (NLP) module of my MSc at UCL. We explored a novel architecture for a contextual bandit with natural language context.