I am a Ph.D. student in robot learning advised by Prof. Jan Peters at the Intelligent Autonomous Systems group at TU Darmstadt. My research interests include robot learning, (inverse) reinforcement learning, imitation learning, and decision-making under partial observability. Currently, my research aims to improve the state of robotic locomotion and manipulation through data-driven reward functions, reinforcement learning, and visual/language inputs.

I received my MSc. in Robotics with cum laude honours from TU Delft where I worked on generative imitation learning techniques under the supervision of Prof. Jens Kober. In the past, I interned in multi-agent robotics at Bosch Research and held positions in graph-based ML at the Wadhwani School of Data Science & AI at IIT Madras.

Email: anish.diwan [at] robot-learning [dot] de
            anish.diwan [at] tu-darmstadt [dot] de

Socials: LinkedIn | GitHub | Notes

news

May 01, 2026	New Paper: Trust Region IRL was accepted at ICML 2026
Apr 16, 2026	I will attend the RL Summer School 2026!
Apr 01, 2025	I started my Ph.D. in robot learning at IAS, TU Darmstadt!
Jan 22, 2025	My MSc. thesis was accepted at ICLR 2025
Sep 19, 2024	I graduated MSc. Robotics from TU Delft with cum laude! 🎓

publications

1. 

   Trust Region Inverse Reinforcement Learning: Explicit Dual Ascent using Local Policy Updates

   Anish Abhijit Diwan, Davide Tateo, Christopher.E Mower, Haitham Bou-Ammar, Jan Peters, and Oleg Arenz

   In International Conference on Machine Learning (ICML), 2026

   Abs arXiv Website

   Inverse reinforcement learning (IRL) is typically formulated as maximizing entropy subject to matching the distribution of expert trajectories. Classical (dual-ascent) IRL guarantees monotonic performance improvement but requires fully solving an RL problem each iteration to compute dual gradients. More recent adversarial methods avoid this cost at the expense of stability and monotonic dual improvement, by directly optimizing the primal problem and using a discriminator to provide rewards. In this work, we bridge the gap between these approaches by enabling monotonic improvement of the reward function and policy without having to fully solve an RL problem at every iteration. Our key theoretical insight is that a trust-region-optimal policy for a reward function update can be globally optimal for a smaller update in the same direction. This smaller update allows us to explicitly optimize the dual objective while only relying on a local search around the current policy. In doing so, our approach avoids the training instabilities of adversarial methods, offers monotonic performance improvement, and learns a reward function in the traditional sense of IRL–one that can be globally optimized to match expert demonstrations. Our proposed algorithm, Trust Region Inverse Reinforcement Learning (TRIRL), outperforms state-of-the-art imitation learning methods across multiple challenging tasks by a factor of 2.4x in terms of aggregate inter-quartile mean, while recovering reward functions that generalize to system dynamics shifts.

2. 

   Noise-conditioned Energy-based Annealed Rewards (NEAR): A Generative Framework for Imitation Learning from Observation

   Anish Abhijit Diwan, Julen Urain, Jens Kober, and Jan Peters

   In International Conference on Learning Representations (ICLR), 2025

   Abs arXiv Website

   This paper introduces a new imitation learning framework based on energy-based generative models capable of learning complex, physics-dependent, robot motion policies through state-only expert motion trajectories. Our algorithm, called Noise-conditioned Energy-based Annealed Rewards (NEAR), constructs several perturbed versions of the expert’s motion data distribution and learns smooth, and well-defined representations of the data distribution’s energy function using denoising score matching. We propose to use these learnt energy functions as reward functions to learn imitation policies via reinforcement learning. We also present a strategy to gradually switch between the learnt energy functions, ensuring that the learnt rewards are always well-defined in the manifold of policy-generated samples. We evaluate our algorithm on complex humanoid tasks such as locomotion and martial arts and compare it with state-only adversarial imitation learning algorithms like Adversarial Motion Priors (AMP). Our framework sidesteps the optimisation challenges of adversarial imitation learning techniques and produces results comparable to AMP in several quantitative metrics across multiple imitation settings.