AI研究論文每日精選

xVerify: Efficient Answer Verifier for Reasoning Model Evaluations

Genius: A Generalizable and Purely Unsupervised Self-Training Framework For Advanced Reasoning

How Instruction and Reasoning Data shape Post-Training: Data Quality through the Lens of Layer-wise Gradients

Seedream 3.0 Technical Report

Heimdall: test-time scaling on the generative verification

Pixel-SAIL: Single Transformer For Pixel-Grounded Understanding

TextArena

Efficient Reasoning Models: A Survey

NormalCrafter: Learning Temporally Consistent Normals from Video Diffusion Priors

DataDecide: How to Predict Best Pretraining Data with Small Experiments

The Scalability of Simplicity: Empirical Analysis of Vision-Language Learning with a Single Transformer

Efficient Process Reward Model Training via Active Learning

DeepMath-103K: A Large-Scale, Challenging, Decontaminated, and Verifiable Mathematical Dataset for Advancing Reasoning

SimpleAR: Pushing the Frontier of Autoregressive Visual Generation through Pretraining, SFT, and RL

RealHarm: A Collection of Real-World Language Model Application Failures

Efficient Generative Model Training via Embedded Representation Warmup

D^2iT: Dynamic Diffusion Transformer for Accurate Image Generation

Efficient Hybrid Language Model Compression through Group-Aware SSM Pruning

A Minimalist Approach to LLM Reasoning: from Rejection Sampling to Reinforce

ReZero: Enhancing LLM search ability by trying one-more-time

VisualPuzzles: Decoupling Multimodal Reasoning Evaluation from Domain Knowledge

AI-University: An LLM-based platform for instructional alignment to scientific classrooms

Diffusion Distillation With Direct Preference Optimization For Efficient 3D LiDAR Scene Completion

PVUW 2025 Challenge Report: Advances in Pixel-level Understanding of Complex Videos in the Wild

Multimodal Long Video Modeling Based on Temporal Dynamic Context

Adaptive Computation Pruning for the Forgetting Transformer

LazyReview A Dataset for Uncovering Lazy Thinking in NLP Peer Reviews

Summarization of Multimodal Presentations with Vision-Language Models: Study of the Effect of Modalities and Structure

Aligning Generative Denoising with Discriminative Objectives Unleashes Diffusion for Visual Perception

Change State Space Models for Remote Sensing Change Detection