AI研究論文每日精選

HtmlRAG: HTML is Better Than Plain Text for Modeling Retrieved Knowledge in RAG Systems

LLaMo: Large Language Model-based Molecular Graph Assistant

Controlling Language and Diffusion Models by Transporting Activations

Adaptive Length Image Tokenization via Recurrent Allocation

DeeR-VLA: Dynamic Inference of Multimodal Large Language Models for Efficient Robot Execution

Sample-Efficient Alignment for LLMs

DreamPolish: Domain Score Distillation With Progressive Geometry Generation

GarVerseLOD: High-Fidelity 3D Garment Reconstruction from a Single In-the-Wild Image using a Dataset with Levels of Details

Inference Optimal VLMs Need Only One Visual Token but Larger Models

Zebra-Llama: A Context-Aware Large Language Model for Democratizing Rare Disease Knowledge

Correlation of Object Detection Performance with Visual Saliency and Depth Estimation