Spread Preference Annotation: Direct Preference Judgment for Efficient LLM Alignment (arXiv, code)

Subtask-Aware Visual Reward Learning from Segmented Demonstrations (arXiv, site)

Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think (arXiv, code, site)

Learning to Contextualize Web Pages for Enhanced Decision Making by LLM Agents

DiffusionGuard: A Robust Defense Against Malicious Diffusion-based Image Editing (arXiv, code, site)

Controllable Blur Data Augmentation Using 3D-Aware Motion Estimation

DiffExp: Efficient Exploration in Reward Fine-tuning for Text-to-Image Diffusion Models (arXiv)

Convergence and Correctness of Max-Product Belief Propagation for Linear Programming

Approximating Spectral Sums of Large-Scale Matrices using Stochastic Chebyshev Approximations

Scheduling Using Interactive Optimization Oracles for Constrained Queueing Networks