Xinsong Feng
  • News
  • Publications
  • Education
  • Teaching
  • Honors & Awards
  • #paper-notes ★
  • #diffusion-model
  • #generative-AI
  • #ML
  • #RL
  • Paper Notes -- Consistency Models

    1/18/2025

    Notes for this paper

  • Paper Notes -- Improved Techniques for Training Consistency Models

    1/13/2025

    Notes for this paper

  • How Log-Likelihood Helps Approximate the True Data Distribution?

    9/19/2024

    Some thoughts about why we use log-likelihood as the loss function when we approximate the true data distribution.

  • Multi-Arm Bandits

    9/18/2024

    Note for Chapter 2 of RL: an introduction

  • Paper Notes - Diffusion Policies creating a Trust Region for Offline Reinforcement Learning

    9/16/2024

    This paper proposed to train a one-step policy guided by a diffusion model via a trust region loss, enabling fast inference without iterative sampling.

  • Paper Notes - Efficient Diffusion Policies for Offline Reinforcement Learning

    9/12/2024

    Notes for this paper.

  • Paper Notes - Diffusion Policies as an Expressive Policy Class for Offline Reinforcement Learning

    9/10/2024

    Notes for this paper.

No posts found matching your search.
1 2
Next

© 2025 Xinsong Feng. All rights reserved.