new

Get trending papers in your email inbox once a day!

Get trending papers in your email inbox!

Daily Papers

byAK and the research community

January

Submitted by

AdinaY

mHC: Manifold-Constrained Hyper-Connections

deepseek-ai

Submitted by

sliuau

GDPO: Group reward-Decoupled Normalization Policy Optimization for Multi-reward RL Optimization

nvidia

Submitted by

Yu2020

Watching, Reasoning, and Searching: A Video Deep Research Benchmark on Open Web for Agentic Video Reasoning

QuantaAlpha

Submitted by

jiaruz2

Agentic Reasoning for Large Language Models

UIUC-CS

University of Illinois at Urbana-Champaign

Submitted by

taesiri

BabyVision: Visual Reasoning Beyond Language

·
29 authors

Submitted by

reign12

STEP3-VL-10B Technical Report

stepfun-ai

Submitted by

weizhoudb

Can LLMs Clean Up Your Mess? A Survey of Application-Ready Data Preparation with LLMs

SJTU

Shanghai Jiao Tong University

Submitted by

Wendy-Fly

Idea2Story: An Automated Pipeline for Transforming Research Concepts into Complete Scientific Narratives

AgentAlphaAGI

Submitted by

taesiri

LongCat-Flash-Thinking-2601 Technical Report

meituan-longcat

Submitted by

taesiri

LTX-2: Efficient Joint Audio-Visual Foundation Model

·
29 authors

Submitted by

xiaochonglinghu

Thinking with Map: Reinforced Parallel Map-Augmented Agent for Geolocalization

alibaba-inc

Submitted by

Yikunb

Your Group-Relative Advantage Is Biased

·
13 authors

7

Submitted by

xiaochonglinghu

Urban Socio-Semantic Segmentation with Vision-Language Reasoning

alibaba-inc

Submitted by

taesiri

Youtu-LLM: Unlocking the Native Agentic Potential for Lightweight Large Language Models

tencent

Submitted by

zhiyuanhucs

Rewarding the Rare: Uniqueness-Aware RL for Creative Problem Solving in LLMs

MIT

Massachusetts Institute of Technology

Submitted by

qiuyuu

Advancing Open-source World Models

robbyant

Submitted by

Yuppie1204

NeoVerse: Enhancing 4D World Model with in-the-wild Monocular Videos

·
6 authors

Submitted by

wukeming11

DeepResearchEval: An Automated Framework for Deep Research Task Construction and Agentic Evaluation

Infinity-AILab

Submitted by

mhjiang0408

daVinci-Dev: Agent-native Mid-training for Software Engineering

GAIR

Submitted by

jasonrqh

AgentDoG: A Diagnostic Guardrail Framework for AI Agent Safety and Security

AI45Research

Submitted by

YanqiDai

Harder Is Better: Boosting Mathematical Reasoning via Difficulty-Aware GRPO and Multi-Aspect Question Reformulation

GD-ML

Submitted by

tristanli

Youtu-Agent: Scaling Agent Productivity with Automated Generation and Hybrid Policy Optimization

tencent

Submitted by

Huacan-Wang

Controlled Self-Evolution for Algorithmic Code Optimization

QuantaAlpha

Submitted by

ssl-asuka

Entropy-Adaptive Fine-Tuning: Resolving Confident Conflicts to Mitigate Forgetting

Submitted by

xiaochonglinghu

Everything in Its Place: Benchmarking Spatial Intelligence of Text-to-Image Models

alibaba-inc

Submitted by

BishopGorov

Improving Multi-step RAG with Hypergraph-based Memory for Long-Context Complex Relational Modeling

tencent

Submitted by

taesiri

Let It Flow: Agentic Crafting on Rock and Roll, Building the ROME Model within an Open Agentic Learning Ecosystem

alibaba-inc

Submitted by

menik1126

MMFormalizer: Multimodal Autoformalization in the Wild

·
14 authors

Submitted by

taesiri

Scaling Embeddings Outperforms Scaling Experts in Language Models

meituan-longcat

Submitted by

haotongl

InfiniDepth: Arbitrary-Resolution and Fine-Grained Depth Estimation with Neural Implicit Fields

zju

Zhejiang University

Submitted by

VentureZJ

MAXS: Meta-Adaptive Exploration with LLM Agents

XianJiaotongUniversity

Xi'an Jiaotong University

Submitted by

rajkumarrawal

Recursive Language Models

MIT

Massachusetts Institute of Technology

Submitted by

taesiri

K-EXAONE Technical Report

LGAI-EXAONE

Submitted by

LutherXD

EvoCUA: Evolving Computer Use Agents via Learning from Scalable Synthetic Experience

meituan

Submitted by

zhiyuanhucs

Collaborative Multi-Agent Test-Time Reinforcement Learning for Reasoning

MIT

Massachusetts Institute of Technology

Submitted by

taesiri

SWE-Pruner: Self-Adaptive Context Pruning for Coding Agents

ByteDance

Submitted by

Bang-UdeM-Mila

Evolving Programmatic Skill Networks

Université de Montréal

Submitted by

daixuancheng

LLM-in-Sandbox Elicits General Agentic Intelligence

MicrosoftResearch

Microsoft Research

Submitted by

reign12

PaCoRe: Learning to Scale Test-Time Compute with Parallel Coordinated Reasoning

stepfun-ai

Submitted by

AmirhoseinGH

Can LLMs Predict Their Own Failures? Self-Awareness via Internal Circuits

UAlberta

University of Alberta

Submitted by

VentureZJ

A^3-Bench: Benchmarking Memory-Driven Scientific Reasoning via Anchor and Attractor Activation

XianJiaotongUniversity

Xi'an Jiaotong University

Submitted by

zichenwen

Innovator-VL: A Multimodal Large Language Model for Scientific Discovery

SJTU

Shanghai Jiao Tong University

Submitted by

Yu2020

MemGovern: Enhancing Code Agents through Learning from Governed Human Experiences

QuantaAlpha

Submitted by

taesiri

FlowAct-R1: Towards Interactive Humanoid Video Generation

ByteDance

4

Submitted by

zawnpn

Being-H0.5: Scaling Human-Centric Robot Learning for Cross-Embodiment Generalization

BeingBeyond

Submitted by

taesiri

Qwen3-TTS Technical Report

Qwen

Submitted by

hzxie

DynamicVLA: A Vision-Language-Action Model for Dynamic Object Manipulation

mmlab-ntu

Submitted by

freesky

HERMES: KV Cache as Hierarchical Memory for Efficient Streaming Video Understanding

OpenMOSS-Team

Submitted by

nzl-thu

The Flexibility Trap: Why Arbitrary Order Limits Reasoning Potential in Diffusion Language Models

Tsinghua-LeapLab

Tsinghua-LeapLab

Submitted by

taesiri

Motion Attribution for Video Generation

nvidia

Submitted by

taesiri

DeepSeek-OCR 2: Visual Causal Flow

deepseek-ai

Submitted by

red-fox-yj

ABC-Bench: Benchmarking Agentic Backend Coding in Real-World Development

OpenMOSS-Team

Submitted by

Minbyul

Solar Open Technical Report

upstage

Submitted by

ScottQu

Dynamic Large Concept Models: Latent Reasoning in an Adaptive Semantic Space

ByteDance-Seed

4

Submitted by

iitolstykh

VIBE: Visual Instruction Based Editor

·
10 authors

Submitted by

itaowe

Advances and Frontiers of LLM-based Issue Resolution in Software Engineering: A Comprehensive Survey

DeepSoftwareAnalytics

Submitted by

ystluffy

Distribution-Aligned Sequence Distillation for Superior Long-CoT Reasoning

Alibaba-Apsara

Alibaba Cloud Apsara Lab

Submitted by

leo1117

NextFlow: Unified Sequential Modeling Activates Multimodal Understanding and Generation

ByteDance

Submitted by

xiaochonglinghu

Taming Hallucinations: Boosting MLLMs' Video Understanding via Counterfactual Video Generation

GD-ML

Submitted by

LHL3341

MMFineReason: Closing the Multimodal Reasoning Gap via Open Data-Centric Methods

SJTU

Shanghai Jiao Tong University

Submitted by

vzl123

RubricHub: A Comprehensive and Highly Discriminative Rubric Dataset via Automated Coarse-to-Fine Generation

·
9 authors

Submitted by

taesiri

Ministral 3

mistralai

2

Submitted by

ngc7293

MOSS Transcribe Diarize: Accurate Transcription with Speaker Diarization

OpenMOSS-Team

Submitted by

LightChen2333

The Molecular Structure of Thought: Mapping the Topology of Long Chain-of-Thought Reasoning

ByteDance

4

Submitted by

Yu2020

KnowMe-Bench: Benchmarking Person Understanding for Lifelong Digital Companions

QuantaAlpha

Submitted by

thenlper

Qwen3-VL-Embedding and Qwen3-VL-Reranker: A Unified Framework for State-of-the-Art Multimodal Retrieval and Ranking

Qwen

Submitted by

adwardlee

Toward Efficient Agents: Memory, Tool learning, and Planning

Submitted by

taesiri

Avatar Forcing: Real-Time Interactive Head Avatar Generation for Natural Conversation

·
5 authors

Submitted by

zptu

The Script is All You Need: An Agentic Framework for Long-Horizon Dialogue-to-Cinematic Video Generation

Tencent-Hunyuan

Tencent Hunyuan

5

Submitted by

bytetriper

Scaling Text-to-Image Diffusion Transformers with Representation Autoencoders

nyu-visionx

Submitted by

jasper0314-huang

Fast-ThinkAct: Efficient Vision-Language-Action Reasoning via Verbalizable Latent Planning

nvidia

2

Submitted by

Qiuchen-Wang

ArenaRL: Scaling RL for Open-Ended Agents via Tournament-based Relative Ranking

Alibaba-NLP

Submitted by

eldad929

CaricatureGS: Exaggerating 3D Gaussian Splatting Faces With Gaussian Curvature

Technion

Technion Israel institute of technology

3

Submitted by

Facico

Stable-DiffCoder: Pushing the Frontier of Code Diffusion Large Language Model

ByteDance-Seed

Submitted by

Mqleet

Paper2Rebuttal: A Multi-Agent Framework for Transparent Author Response Assistance

Submitted by

Minbyul

User-Oriented Multi-Turn Dialogue Generation with Tool Use at scale

upstage

Submitted by

XuGuo699

DreamID-V:Bridging the Image-to-Video Gap for High-Fidelity Face Swapping via Diffusion Transformer

ByteDance

Submitted by

donghao-zhou

MMDeepResearch-Bench: A Benchmark for Multimodal Deep Research Agents

·
15 authors

Submitted by

Spico

DiffThinker: Towards Generative Multimodal Reasoning with Diffusion Models

·
6 authors

Submitted by

xxxllz

OCRVerse: Towards Holistic OCR in End-to-End Vision-Language Models

·
10 authors

Submitted by

Lin-Chen

UniCorn: Towards Self-Improving Unified Multimodal Models through Self-Generated Supervision

ustc-community

University of Science and Technology of China

Submitted by

zkcys001

A Pragmatic VLA Foundation Model

robbyant

Submitted by

Dongchao

HeartMuLa: A Family of Open Sourced Music Foundation Models

·
28 authors

Submitted by

hitsmy

AdaReasoner: Dynamic Tool Orchestration for Iterative Visual Reasoning

Fudan-University

Fudan University

Submitted by

PangzeCheung

OmniTransfer: All-in-one Framework for Spatio-temporal Video Transfer

ByteDance

Submitted by

williamium

Think3D: Thinking with Space for Spatial Reasoning

·
12 authors

Submitted by

Blue-Giant

Beyond Static Tools: Test-Time Tool Evolution for Scientific Reasoning

Submitted by

NeoZ123

Chaining the Evidence: Robust Reinforcement Learning for Deep Search Agents with Citation-Aware Rubric Rewards

zai-org

Submitted by

yulunliu

RL-AWB: Deep Reinforcement Learning for Auto White Balance Correction in Low-Light Night-time Scenes

·
4 authors

Submitted by

jonhue

Reinforcement Learning via Self-Distillation

lasgroup

LAS @ ETH Zurich

Submitted by

EilamSha

The Poisoned Apple Effect: Strategic Manipulation of Mediated Markets via Technology Expansion of AI Agents

Technion

Technion Israel institute of technology

4

Submitted by

taesiri

X-Coder: Advancing Competitive Programming with Fully Synthetic Tasks, Solutions, and Tests

MicrosoftResearch

Microsoft Research

Submitted by

taesiri

NitroGen: An Open Foundation Model for Generalist Gaming Agents

nvidia

Submitted by

AliBehrouz

Nested Learning: The Illusion of Deep Learning Architectures

·
4 authors

8

Submitted by

vinid

Learning to Discover at Test Time

StanfordUniversity

Stanford University

Submitted by

taesiri

Rethinking Video Generation Model for the Embodied World

ByteDance-Seed

Submitted by

Jinyang23

Atlas: Orchestrating Heterogeneous Models and Tools for Multi-Domain Complex Reasoning

·
8 authors

2

Submitted by

NasirzadehMoh

A unified framework for detecting point and collective anomalies in operating system logs via collaborative transformers

alarmif

Submitted by

JingweiZuo

Learnable Multipliers: Freeing the Scale of Language Model Matrix Layers

tiiuae

Technology Innovation Institute

3

Submitted by

Knykny

Teaching Models to Teach Themselves: Reasoning at the Edge of Learnability

3

Submitted by

LXT

SAMTok: Representing Any Mask with Two Words

ByteDance

Submitted by

KevinQHLin

ShowUI-π: Flow-based Generative Models as GUI Dexterous Hands

showlab

Submitted by

FetchFortune

ConceptMoE: Adaptive Token-to-Concept Compression for Implicit Compute Allocation

ByteDance-Seed

Submitted by

LHL3341

Scientific Image Synthesis: Benchmarking, Methodologies, and Downstream Utility

SJTU

Shanghai Jiao Tong University

Submitted by

taesiri

EvoFSM: Controllable Self-Evolution for Deep Research with Finite State Machines

QuantaAlpha