KI-Forschungspapiere Täglich

Long-Context Autoregressive Video Modeling with Next-Frame Prediction

Scaling Vision Pre-Training to 4K Resolution

Inference-Time Scaling for Flow Models via Stochastic Generation and Rollover Budget Forcing

Exploring Hallucination of Large Multimodal Models in Video Understanding: Benchmark, Analysis and Mitigation

CoMP: Continual Multimodal Pre-training for Vision Foundation Models

Think Twice: Enhancing LLM Reasoning by Scaling Multi-round Test-time Thinking

Spot the Fake: Large Multimodal Model-Based Synthetic Image Detection with Artifact Explanation

MDocAgent: A Multi-Modal Multi-Agent Framework for Document Understanding

ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learning

CoLLM: A Large Language Model for Composed Image Retrieval

WikiAutoGen: Towards Multi-Modal Wikipedia-Style Article Generation

Latent Space Super-Resolution for Higher-Resolution Image Generation with Diffusion Models

FullDiT: Multi-Task Video Generative Foundation Model with Full Attention

DiffPortrait360: Consistent Portrait Diffusion for 360 View Synthesis

FirePlace: Geometric Refinements of LLM Common Sense Reasoning for 3D Object Placement

LookAhead Tuning: Safer Language Models via Partial Answer Previews

PhysTwin: Physics-Informed Reconstruction and Simulation of Deformable Objects from Videos

Efficient Model Development through Fine-tuning Transfer

FRESA:Feedforward Reconstruction of Personalized Skinned Avatars from Few Images

xKV: Cross-Layer SVD for KV-Cache Compression

Gumbel-Softmax Flow Matching with Straight-Through Guidance for Controllable Biological Sequence Generation

Strong Baseline: Multi-UAV Tracking via YOLOv12 with BoT-SORT-ReID

When Words Outperform Vision: VLMs Can Self-Improve Via Text-Only Training For Human-Centered Decision Making

LLaVAction: evaluating and training multi-modal large language models for action recognition

OpenCity3D: What do Vision-Language Models know about Urban Environments?

Towards a Unified Copernicus Foundation Model for Earth Vision

Can Vision-Language Models Answer Face to Face Questions in the Real-World?

Overcoming Vocabulary Mismatch: Vocabulary-agnostic Teacher Guided Language Modeling

Frequency Dynamic Convolution for Dense Image Prediction

Any6D: Model-free 6D Pose Estimation of Novel Objects

LPOSS: Label Propagation Over Patches and Pixels for Open-vocabulary Semantic Segmentation

ST-VLM: Kinematic Instruction Tuning for Spatio-Temporal Reasoning in Vision-Language Models

Co-SemDepth: Fast Joint Semantic Segmentation and Depth Estimation on Aerial Images