AI 연구 논문 데일리

Web-Shepherd: Advancing PRMs for Reinforcing Web Agents

MMaDA: Multimodal Large Diffusion Language Models

Scaling Law for Quantization-Aware Training

UniVG-R1: Reasoning Guided Universal Visual Grounding with Reinforcement Learning

Diffusion vs. Autoregressive Language Models: A Text Embedding Perspective

Efficient Agent Training for Computer Use

This Time is Different: An Observability Perspective on Time Series Foundation Models

Learn to Reason Efficiently with Adaptive Length-based Reward Shaping

Constructing a 3D Town from a Single Image

When to Continue Thinking: Adaptive Thinking Mode Switching for Efficient Reasoning

Vid2World: Crafting Video Diffusion Models to Interactive World Models

lmgame-Bench: How Good are LLMs at Playing Games?

VerifyBench: Benchmarking Reference-based Reward Systems for Large Language Models

Deliberation on Priors: Trustworthy Reasoning of Large Language Models on Knowledge Graphs

IA-T2I: Internet-Augmented Text-to-Image Generation

Learning to Reason via Mixture-of-Thought for Logical Reasoning

dKV-Cache: The Cache for Diffusion Language Models

Be Careful When Fine-tuning On Open-Source LLMs: Your Fine-tuning Data Could Be Secretly Stolen!

How Should We Enhance the Safety of Large Reasoning Models: An Empirical Study

Soft Thinking: Unlocking the Reasoning Potential of LLMs in Continuous Concept Space

RLVR-World: Training World Models with Reinforcement Learning

BARREL: Boundary-Aware Reasoning for Factual and Reliable LRMs

ConvSearch-R1: Enhancing Query Reformulation for Conversational Search with Reasoning via Reinforcement Learning

Text Generation Beyond Discrete Token Sampling

Evaluate Bias without Manual Test Sets: A Concept Representation Perspective for LLMs

AutoMat: Enabling Automated Crystal Structure Reconstruction from Microscopy via Agentic Tool Use

DiCo: Revitalizing ConvNets for Scalable and Efficient Diffusion Modeling

Audio Jailbreak: An Open Comprehensive Benchmark for Jailbreaking Large Audio-Language Models

VARD: Efficient and Dense Fine-Tuning for Diffusion Models with Value-based RL

PiFlow: Principle-aware Scientific Discovery with Multi-Agent Collaboration

RL Tango: Reinforcing Generator and Verifier Together for Language Reasoning

WebNovelBench: Placing LLM Novelists on the Web Novel Distribution

Prior Prompt Engineering for Reinforcement Fine-Tuning

BLEUBERI: BLEU is a surprisingly effective reward for instruction following

Streamline Without Sacrifice - Squeeze out Computation Redundancy in LMM

Scaling and Enhancing LLM-based AVSR: A Sparse Mixture of Projectors Approach

BanditSpec: Adaptive Speculative Decoding via Bandit Algorithms

The Unreasonable Effectiveness of Entropy Minimization in LLM Reasoning

MultiHal: Multilingual Dataset for Knowledge-Graph Grounded Evaluation of LLM Hallucinations

HumaniBench: A Human-Centric Framework for Large Multimodal Models Evaluation

Language Specific Knowledge: Do Models Know Better in X than in English?

In-Context Learning Boosts Speech Recognition via Human-like Adaptation to Speakers and Language Varieties