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

Aug 4

Submitted by

AaronZ345

SwanTale: Unified Multi-Speaker Speech and Audio Generation for Instruct and Zero-Shot Tasks

ByteDance

7

Submitted by

seashell11

LongHorizon-Harness: Advancing Long-Horizon Agents for Real-World Tasks

·
8 authors

Submitted by

shenjunhao

Progressive Agent Skill Generation via Reinforcement Learning

CUHKDBGROUP

The Chinese University of Hong Kong - Database Group

Submitted by

Uanu

DAPD: Dual-Anchored Policy Distillation

Shanghai-AI-Laboratory

Shanghai AI Laboratory

2

Submitted by

zhangkangning

VAD: Attributing Visual Evidence for Target Reconstruction in Multimodal On-Policy Distillation

·
12 authors

Submitted by

Yuppie1204

WorldExam: Benchmarking World Models from Apparent Appearance to Inherent Reactivity

·
16 authors

Submitted by

songtingyu

UEmbed: Unified Sparse and Dense Multimodal Embeddings

Alibaba-NLP

Submitted by

Artemis0430

SKT: Skill-Use Training at Scale via Verified Synthetic Data Generation

Shanghai-AI-Laboratory

Shanghai AI Laboratory

1

Submitted by

zhemchuzhnikov

CADENA: Stepwise CAD Reverse Engineering

·
12 authors

Submitted by

hzxllll

Deferred Exposure of Future Trajectories for Verifiable Reasoning in Autonomous Driving VLMs

·
9 authors

Submitted by

Trae1ounG

SWE-Touch: Benchmarking Coding Agents When Users Touch the Code

CASIA

Chinese Academic of Science Institute of Automation

Submitted by

Sylvest

WCM: A World Critic Model for Vision-Language-Action Reinforcement Learning

OpenMOSS-Team

Submitted by

zmmm

Motion Beyond Morphology: Bootstrapping Cross-Category Motion Transfer from Abstract Motion Representations

KlingTeam

Submitted by

yzx010

GradCuit: Credit-Assigned Gradient Flow Enables Robust and Interpretable Test-Time Latent Reasoning

·
7 authors

Submitted by

levondang

Roomer: Reflective Object-Grounded Model Editing and Repair for 3D Indoor Layout Synthesis

cn-scut

South China University of Technology

Submitted by

taesiri

DiffusionGemma Technical Report

deepmind

Submitted by

softmax

GPTQ-2D: Cubic-Time Two-Sided Adaptive Rounding

ISTA-DASLab

IST Austria Distributed Algorithms and Systems Lab

1

Submitted by

Halcyon-Zhang

DeepVoyager-VL: Incentivizing Vision-in-the-Loop Search for Long-Horizon Multimodal Agents

PekingUniversity

Peking University

Submitted by

higokri

3DZip: Spatial-Aware Feature Diversity-Guided Token Compression for 3D Question Answering

·
2 authors

Submitted by

z-rx

LeapTalk: Breaking the Latency-Quality Trade-off in Talking Head Generation

Shanghai-Jiao-Tong-University-SAI

Shanghai Jiao Tong University SAI

Submitted by

levondang

StyleForge: Indoor Furniture Styling by Counterfactual Reasoning in a Hypergraph Field

cn-scut

South China University of Technology

Submitted by

dts1347

DreamTraj: Generating 6-DoF Object Trajectories by Reading Unrendered Video Diffusion Latents

SUSTech

Southern university of science and technology

1

Submitted by

choucsan

Poplar: A Scalable Pipeline for Human-Centric Image Dataset Synthesis

·
1 authors

Submitted by

vernontoh

ScrambleToolBench: Agents Search Exhaustively Even When Their Own Map Points to the Next Step

declare-lab

Deep Cognition and Language Research (DeCLaRe) Lab

Submitted by

ZiangWu

Relax Within, Balance Across: Geometry-Guided Load Balancing for Vision-Language Mixture-of-Experts

PekingUniversity

Peking University

Submitted by

6lyc

RecHarness: A Bandit-Routed Agentic Harness for Self-Evolving Recommender Systems

kuaishou technology

Submitted by

Harsh1729

Model or Harness? An Interaction-Centric Taxonomy for Localizing Agent Failures

ScaleAI

Submitted by

jdsouza

ICDAR 2026 Competition on Information Extraction from Atomic Layer Deposition/Etching (ALD/E) Scientific Figures

SciKnowOrg

Scientific Knowledge Organization

Submitted by

nielsr

A Frozen Pixel-Space Diffusion Model Can Guide Itself with Its Own Samples

nanyang-technological-university-singapore

Nanyang Technological University Singapore

Submitted by

oklenAI

Compute Globally, Materialize Locally: The Memory Contract of Sparse Event-KV

·
2 authors

Submitted by

gtano

GEOID-Flood: A Large-Scale Multi-Modal Benchmark Dataset for Flood Segmentation

links-ads

LINKS - AI, Data & Space