KI-Forschungspapiere Täglich

Can large language models explore in-context?

LLM2LLM: Boosting LLMs with Novel Iterative Data Enhancement

InternVideo2: Scaling Video Foundation Models for Multimodal Video Understanding

Champ: Controllable and Consistent Human Image Animation with 3D Parametric Guidance

ThemeStation: Generating Theme-Aware 3D Assets from Few Exemplars

VidLA: Video-Language Alignment at Scale

SiMBA: Simplified Mamba-Based Architecture for Vision and Multivariate Time series

DragAPart: Learning a Part-Level Motion Prior for Articulated Objects

FollowIR: Evaluating and Teaching Information Retrieval Models to Follow Instructions

AllHands: Ask Me Anything on Large-scale Verbatim Feedback via Large Language Models

LATTE3D: Large-scale Amortized Text-To-Enhanced3D Synthesis

Compiler generated feedback for Large Language Models