Reinforcement Learning on the Discrete Composition Channel of a Crystal Generator: Validated Gains and Reward Hacking

arXiv:2610.03880 · cs.LG, cond-mat.mtrl-sci · Submitted 2026-10-02 · Read on arXiv

cs.LG, cond-mat.mtrl-sci

Submitted: 2026-10-02

Updated: 2026-10-02

Code: https://github.com/paprakash/OMatGRPO

Terminology

• Reinforcement Learning — 1× in this paper · explained in 4 episodes
• Reward Hacking — 1× in this paper · explained in 8 episodes

Sources

• Discovery of Spin-Crossover Candidates with Equivariant Graph Neural Networks and Relevance-Based Classification
• Stochastic Interpolants: A Unifying Framework for Flows and Diffusions
• Concrete Problems in AI Safety
• LeMat-GenBench: A Unified Evaluation Framework for Crystal Generative Models
• Training Diffusion Models with Reinforcement Learning
• Generative Flows on Discrete State-Spaces: Enabling Multimodal Flows with Applications to Protein Co-Design
• Reinforcement Fine-Tuning for Materials Design
• Accelerating inverse materials design using generative diffusion models with reinforcement learning
• DPOK: Reinforcement Learning for Fine-tuning Text-to-Image Diffusion Models
• Scaling Laws for Reward Model Overoptimization
• Discrete Flow Matching
• CrystalGym: A New Benchmark for Materials Discovery Using Reinforcement Learning
• KL-Regularized Reinforcement Learning is Designed to Mode Collapse
• Crystal-GFN: sampling crystals with desirable properties and constraints
• Classifier-Free Diffusion Guidance
• Open Materials Generation with Inference-Time Reinforcement Learning
• Open Materials Generation with Stochastic Interpolants
• Symmetry- and Property-Aware Crystal Generation with Reinforcement Learning for Inverse Materials Design
• Crystal Structure Prediction by Joint Equivariant Diffusion
• Deep Reinforcement Learning for Inverse Inorganic Materials Design

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks