AI研究論文每日精選

Byte Latent Transformer: Patches Scale Better Than Tokens

Evaluation Agent: Efficient and Promptable Evaluation Framework for Visual Generative Models

BrushEdit: All-In-One Image Inpainting and Editing

RetroLLM: Empowering Large Language Models to Retrieve Fine-grained Evidence within Generation

Smaller Language Models Are Better Instruction Evolvers

ColorFlow: Retrieval-Augmented Image Sequence Colorization

Causal Diffusion Transformers for Generative Modeling

SPaR: Self-Play with Tree-Search Refinement to Improve Instruction-Following in Large Language Models

Wonderland: Navigating 3D Scenes from a Single Image

GaussianProperty: Integrating Physical Properties to 3D Gaussians with LMMs

VividFace: A Diffusion-Based Hybrid Framework for High-Fidelity Video Face Swapping

IDArb: Intrinsic Decomposition for Arbitrary Number of Input Views and Illuminations

StrandHead: Text to Strand-Disentangled 3D Head Avatars Using Hair Geometric Priors

SepLLM: Accelerate Large Language Models by Compressing One Segment into One Separator

Emma-X: An Embodied Multimodal Action Model with Grounded Chain of Thought and Look-ahead Spatial Reasoning

The Open Source Advantage in Large Language Models (LLMs)

SplineGS: Robust Motion-Adaptive Spline for Real-Time Dynamic 3D Gaussians from Monocular Video

DynamicScaler: Seamless and Scalable Video Generation for Panoramic Scenes

Wonderful Matrices: Combining for a More Efficient and Effective Foundation Model Architecture

TidyBot++: An Open-Source Holonomic Mobile Manipulator for Robot Learning

MOVIS: Enhancing Multi-Object Novel View Synthesis for Indoor Scenes

GeoX: Geometric Problem Solving Through Unified Formalized Vision-Language Pre-training

MaxInfoRL: Boosting exploration in reinforcement learning through information gain maximization

Whisper-GPT: A Hybrid Representation Audio Large Language Model

Reliable, Reproducible, and Really Fast Leaderboards with Evalica

Just a Simple Transformation is Enough for Data Protection in Vertical Federated Learning

Nearly Zero-Cost Protection Against Mimicry by Personalized Diffusion Models

RLDG: Robotic Generalist Policy Distillation via Reinforcement Learning