ArXiv TLDR

Arun Ross

2 papers · Latest:

Computer Vision

S2H-DPO: Hardness-Aware Preference Optimization for Vision-Language Models

S2H-DPO introduces a Simple-to-Hard learning framework to improve Vision-Language Models' multi-image reasoning by constructing hierarchical preference data.

2604.18512
Computer Vision

Are Face Embeddings Compatible Across Deep Neural Network Models?

This paper finds that face embeddings from different deep neural network models are surprisingly compatible and can be aligned with simple transformations.

2604.07282

📬 Weekly AI Paper Digest

Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.