Aditya Mohan

Researcher and PhD student at the Institute of Artificial Intellifence, Hannover, working under the supervision of Prof. Marius Lindauer

prof-pic.jpg

Sequential decision-making problems in the real world often come with abundant information that is typically underutilized in Reinforcement Learning (RL). My research focuses on leveraging this information to enhance RL techniques. Specifically, I aim to explore how RL agents can utilize task structures to improve sample efficiency and generalization. In doing so, my work intersects various RL domains, including Contextual RL, Automated RL, Meta-RL, and state abstractions.

I completed my Master’s in Autonomous Systems from Technical University of Berlin and EURECOM, where my thesis focused on ad-hoc cooperation in Hanabi. Additionally, I have worked on problems in Robotics, Reinforcement Learning, and Meta-Learning in both single-agent and multi-agent settings.

Beyond my research, I have a passion for singing, playing piano and guitar, composing music, dancing, and cooking. If you’d like to chat about research, music, food, or anything else, don’t hesitate to book a slot in my calender

news

Apr 3, 2024 Our paper Structure in Deep Reinforcement Learning: A Survey and Open Problems has been published in the Journal of Artificial Intelligence Research (JAIR). Super thankful to my coauthors Amy Zhang and Marius Lindauer!
Mar 21, 2024 Our poster “Instance Selection for Dynamic Algorithm Configuration with Reinforcement Learning: Improving Generalization” has been accepted to The Genetic and Evolutionary Computation Conference (GECCO 2024). Do come by!
Feb 12, 2024 Our paper AutoML in the Age of Large Language Models: Current Challenges, Future Opportunities and Risks has been accepted to the journal on Transactions on Machine Learning Research (TMLR). Check it out if you are interested in learning how AutoML and LLMs compliment each other!
Jan 11, 2024 Raghu Rajan, Theresa Eimer, Andre Bidenkapp and I have written a blog post on the AutoRL blog about the year 2023 in AutoRL research. You can read it here.
Sep 13, 2023 I will be attending EWRL 2023 and presenting the following papers: Additional, Theresa Eimer will present our work Contextualize Me – The Case for Context in Reinforcement Learning
May 31, 2023 Thrilled to have the following two papers accepted at the AutoML Conference 2023 Check out our posters in Berlin!

selected publications

  1. Structure in Deep Reinforcement Learning: A survey and open problems
    Aditya MohanAmy Zhang, and Marius Lindauer
    Journal of Artificial Intelligence Research (JAIR), 2024
  2. AutoRL Hyperparameter Landscapes
    Aditya MohanCarolin Benjamins, Konrad Wienecke, Alexander Dockhorn, and Marius Lindauer
    Proceedings of the Second International Conference on Automated Machine Learning (AutoML), 2023
  3. Contextualize Me - The Case for Context in Reinforcement Learning
    Carolin BenjaminsTheresa EimerF. SchubertAditya Mohan, Andre Biedenkapp, Bodo RosenhahnFrank Hutter, and Marius Lindauer
    Transactions on Machine Learning Research (TMLR), 2023