Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..

Adversarial Diffusion for Robust Reinforcement Learning

Authors: Daniele Foffano, Alessio Russo, Alexandre Proutiere

NeurIPS 2025 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Empirical results across standard benchmarks show that AD-RRL achieves superior robustness and performance compared to existing robust RL methods.
Researcher Affiliation Academia Daniele Foffano Division of Decision and Control Systems KTH, Royal Institute of Technology EMAIL Alessio Russo Faculty of Computing and Data Sciences Boston University EMAIL Alexandre Proutiere Division of Decision and Control Systems KTH, Royal Institute of Technology EMAIL
Pseudocode Yes Algorithm 1 Adversarial Diffusion for Robust Reinforcement Learning (AD-RRL) ... Algorithm 2 Adversarial Diffusion Trajectory Sampling ... Algorithm 3 Diffusion model training
Open Source Code Yes The official implementation for AD-RRL can be found on Git Hub: https://github.com/ danielefoffano/AD-RRL
Open Datasets Yes Our experiments are conducted on several optimal control tasks from the Mu Jo Co suite: Inverted Pendulum, Reacher, Hopper, Half Cheetah, and Walker.
Dataset Splits No All agents are trained in the default Mu Jo Co/Open AI Gym environment (fixed physics), for 1.5M steps. ... At test time, we alter key physics-related parameters and assess the agent s performance against both robust and non-robust baselines.
Hardware Specification Yes The training of AD-RRL and Polygrad was performed on three different machines. On a cluster node with one A100 GPU, Icelake CPU and 256 GB of RAM. The remaining model-free baselines were trained on a laptop with an Intel i7-1185G7 CPU, Mesa Intel Xe Graphics GPU and 32 GB of RAM.
Software Dependencies No The implementation of Poly GRAD was taken from the respective github repository [38]. The same was done for CPPO and M2TD3... For TRPO and PPO we used the implementation from Stable-Baselines3 [36].
Experiment Setup Yes Table 2: Hyperparameters for A2C training. ... Table 3: Hyperparameters for adversarial diffusion training.