LASER: Latent Space Adjoint Matching for Support-Constrained Entropy-Regularized Offline RL

arXiv:2610.08989 · cs.LG, cs.AI, cs.RO, math.OC, stat.ML · Submitted 2026-10-06 · Read on arXiv

cs.LG, cs.AI, cs.RO, math.OC, stat.ML

Submitted: 2026-10-06

Updated: 2026-10-14

Comments: 29 pages, 16 figures. Accepted at NeurIPS 2026

Project page: https://mit-realm.github.io/laser

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Latent Space — 1× in this paper · explained in 7 episodes

Sources

• Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems
• Advantage-Weighted Regression: Simple and Scalable Off-Policy Reinforcement Learning
• Behavior Regularized Offline Reinforcement Learning
• Steering Your Diffusion Policy with Latent Space Reinforcement Learning
• Building Normalizing Flows with Stochastic Interpolants
• Adjoint Matching: Fine-tuning Flow and Diffusion Generative Models with Memoryless Stochastic Optimal Control
• Score-Based Generative Modeling through Stochastic Differential Equations
• IDQL: Implicit Q-Learning as an Actor-Critic Method with Diffusion Policies
• EXPO: Stable Reinforcement Learning with Expressive Policies
• floq: Training Critics via Flow-Matching for Scaling Compute in Value-Based RL
• Way Off-Policy Batch Deep Reinforcement Learning of Implicit Human Preferences in Dialog
• Keep Doing What Worked: Behavioral Modelling Priors for Offline Reinforcement Learning
• AWAC: Accelerating Online Reinforcement Learning with Offline Datasets
• Imitating Human Behaviour with Diffusion Models
• Fisher Decorator: Refining Flow Policy via a Local Transport Map
• Classifier-Free Diffusion Guidance
• Diffusion Guidance Is a Controllable Policy Improvement Operator
• Gaussian Error Linear Units (GELUs)
• Layer Normalization

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