Claas A. Voelcker

PhD Candidate at UofT, RL researcher focused on too many things, he/him, 🏳️‍🌈 🤖 🧙

prof_pic.jpg

W1140-108 College Street

SR Innovation Campus

Toronto, ON

M5G 0C6

I am currently looking for post-doc opportunities!

I am a PhD student in Reinforcement and Machine Learning at the University of Toronto and the Vector Institute, supervised by Profs. Amir-massoud Farahmand and Igor Gilitschenski.

My research focuses on model based reinforcement learning and closing the gap between learning acurate models for future predictions and learning high performing models for planning. I am interested in using techniques for representation and world model learning to stablize notoriously brittle Deep Reinforcement Learning approaches. Finally, I like thinking about how we can do better science in RL by thinking about what problems we should be benchmarking our exciting advances on.

Originally from Germany, I received a Bachelor and Master degree from the University of Darmstadt with Honors. There, I had the great pleasure to be supervised and mentored by Profs. Kristian Kersting and Jan Peters.

I am proud to serve as a core organizer for Queer in AI, where I help promote the interests of queer researchers and practitioners at AI /ML conferences and in the wider community.

news

Oct 13, 2024 I made a new website!

selected publications

2024

  1. paper_dissecting.png
    Dissecting Deep RL with High Update Ratios: Combatting Value Overestimation and Divergence
    Marcel Hussing, Claas A. Voelcker, Igor Gilitschenski, Amir-massoud Farahmand, and Eric Eaton
    Reinforcement Learning Conference, Aug 2024
  2. paper_understanding.png
    When does Self-Prediction help? Understanding Auxiliary Tasks in Reinforcement Learning
    Claas A. Voelcker, Tyler Kastner, Igor Gilitschenski, and Amir-massoud Farahmand
    Reinforcement Learning Conference, Aug 2024
  3. paper_mad.png
    MAD-TD: Model-Augmented Data stabilizes High Update Ratio RL
    Claas A. Voelcker, Marcel Hussing, Eric Eaton, Amir-massoud Farahmand, and Igor Gilitschenski
    under review, Oct 2024

2023

  1. paper_lambda.png
    λ-AC: Learning latent decision-aware models for reinforcement learning in continuous state-spaces
    Claas A. Voelcker, Arash Ahmadian, Romina Abachi, Igor Gilitschenski, and Amir-massoud Farahmand
    arXiv preprint arXiv:2306.17366, Nov 2023

2022

  1. paper_vagram.png
    Value Gradient weighted Model-Based Reinforcement Learning
    Claas A. Voelcker, Victor Liao, Animesh Garg, and Amir-massoud Farahmand
    In International Conference on Learning Representations, Apr 2022