AI研究論文每日精選

R1-Onevision: Advancing Generalized Multimodal Reasoning through Cross-Modal Formalization

Autoregressive Image Generation with Randomized Parallel Decoding

Open-Sora 2.0: Training a Commercial-Level Video Generation Model in $200k

On the Limitations of Vision-Language Models in Understanding Image Transforms

Transformers without Normalization

PerCoV2: Improved Ultra-Low Bit-Rate Perceptual Image Compression with Implicit Hierarchical Masked Image Modeling

GroundingSuite: Measuring Complex Multi-Granular Pixel Grounding

Studying Classifier(-Free) Guidance From a Classifier-Centric Perspective

Quantization for OpenAI's Whisper Models: A Comparative Analysis

MinorBench: A hand-built benchmark for content-based risks for children

TruthPrInt: Mitigating LVLM Object Hallucination Via Latent Truthful-Guided Pre-Intervention

World Modeling Makes a Better Planner: Dual Preference Optimization for Embodied Task Planning

OmniPaint: Mastering Object-Oriented Editing via Disentangled Insertion-Removal Inpainting

VisualWebInstruct: Scaling up Multimodal Instruction Data through Web Search

Communication-Efficient Language Model Training Scales Reliably and Robustly: Scaling Laws for DiLoCo

Shifting Long-Context LLMs Research from Input to Output

Silent Branding Attack: Trigger-free Data Poisoning Attack on Text-to-Image Diffusion Models

The Curse of Conditions: Analyzing and Improving Optimal Transport for Conditional Flow-Based Generation

SANA-Sprint: One-Step Diffusion with Continuous-Time Consistency Distillation

CoSTAast: Cost-Sensitive Toolpath Agent for Multi-turn Image Editing

DiT-Air: Revisiting the Efficiency of Diffusion Model Architecture Design in Text to Image Generation

Light-R1: Curriculum SFT, DPO and RL for Long COT from Scratch and Beyond

VisualPRM: An Effective Process Reward Model for Multimodal Reasoning

CoRe^2: Collect, Reflect and Refine to Generate Better and Faster

GoT: Unleashing Reasoning Capability of Multimodal Large Language Model for Visual Generation and Editing

Distilling Diversity and Control in Diffusion Models

CINEMA: Coherent Multi-Subject Video Generation via MLLM-Based Guidance

Discovering Influential Neuron Path in Vision Transformers

UniGoal: Towards Universal Zero-shot Goal-oriented Navigation

Piece it Together: Part-Based Concepting with IP-Priors

ConsisLoRA: Enhancing Content and Style Consistency for LoRA-based Style Transfer

Long Context Tuning for Video Generation

PoseLess: Depth-Free Vision-to-Joint Control via Direct Image Mapping with VLM

"Silent Is Not Actually Silent": An Investigation of Toxicity on Bug Report Discussion

A Frustratingly Simple Yet Highly Effective Attack Baseline: Over 90% Success Rate Against the Strong Black-box Models of GPT-4.5/4o/o1

4D LangSplat: 4D Language Gaussian Splatting via Multimodal Large Language Models

Charting and Navigating Hugging Face's Model Atlas

Do I look like a `cat.n.01` to you? A Taxonomy Image Generation Benchmark

New Trends for Modern Machine Translation with Large Reasoning Models