AI Papers - 2026-04-11

DailyArxiv - AI Research Podcast · Apr 11, 2026 · 13m

Today's papers: - Emotion Concepts and their Function in a Large Language Model: https://arxiv.org/abs/2604.07729v1 - Small Vision-Language Models are Smart Compressors for Long Video Understanding: https://arxiv.org/abs/2604.08120v1 - PokeGym: A Visually-Driven Long-Horizon Benchmark for Vision-Language Models: https://arxiv.org/abs/2604.08340v1 - Uni-ViGU: Towards Unified Video Generation and Understanding via A Diffusion-Based Video Generator: https://arxiv.org/abs/2604.08121v1 - Revise: A Framework for Revising OCRed text in Practical Information Systems with Data Contamination Strategy…