KI-Forschungspapiere Täglich

AceMath: Advancing Frontier Math Reasoning with Post-Training and Reward Modeling

DateLogicQA: Benchmarking Temporal Biases in Large Language Models

Progressive Multimodal Reasoning via Active Retrieval

TOMG-Bench: Evaluating LLMs on Text-based Open Molecule Generation

Affordance-Aware Object Insertion via Mask-Aware Dual Diffusion

Qwen2.5 Technical Report

How to Synthesize Text Data without Model Collapse?

Descriptive Caption Enhancement with Visual Specialists for Multimodal Perception

AV-Link: Temporally-Aligned Diffusion Features for Cross-Modal Audio-Video Generation

LeviTor: 3D Trajectory Oriented Image-to-Video Synthesis

DI-PCG: Diffusion-based Efficient Inverse Procedural Content Generation for High-quality 3D Asset Creation

Move-in-2D: 2D-Conditioned Human Motion Generation

UIP2P: Unsupervised Instruction-based Image Editing via Cycle Edit Consistency

MegaPairs: Massive Data Synthesis For Universal Multimodal Retrieval

LongBench v2: Towards Deeper Understanding and Reasoning on Realistic Long-context Multitasks

PixelMan: Consistent Object Editing with Diffusion Models via Pixel Manipulation and Generation

Flowing from Words to Pixels: A Framework for Cross-Modality Evolution