Discrete MeanFlow: One-Step Generation via Conditional Transition Kernels

arXiv:2605.12805 · cs.LG, cs.AI · Submitted 2026-05-12 · Read on arXiv

cs.LG, cs.AI

Submitted: 2026-05-12

Updated: 2026-10-12

Terminology

• Flow Matching — 1× in this paper · explained in 10 episodes

Sources

• One-Step Flow Policy Mirror Descent
• Improved Mean Flows: On the Challenges of Fastforward Generative Models
• Flow Matching for Offline Reinforcement Learning with Discrete Actions
• MeanAudio: Fast and Faithful Text-to-Audio Generation with Mean Flows
• Flow Matching Guide and Code
• Rectified Flow: A Marginal Preserving Approach to Optimal Transport
• AlphaFlow: Understanding and Improving MeanFlow Models

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