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

Apr 23

Submitted by

taesiri

LLaDA2.0-Uni: Unifying Multimodal Understanding and Generation with Diffusion Large Language Model

inclusionAI

Submitted by

iieycx

Near-Future Policy Optimization

·
9 authors

Submitted by

KID-22

DR-Venus: Towards Frontier Edge-Scale Deep Research Agents with Only 10K Open Data

inclusionAI

Submitted by

fdu-lcz

Reward Hacking in the Era of Large Models: Mechanisms, Emergent Misalignment, Challenges

Fudan-University

Fudan University

Submitted by

HyeonwooKim

DeVI: Physics-based Dexterous Human-Object Interaction via Synthetic Video Imitation

snuvclab

Visual Computing Lab

Submitted by

Z-MU-Z

Exploring Spatial Intelligence from a Generative Perspective

·
12 authors

Submitted by

JinChengRen

A Self-Evolving Framework for Efficient Terminal Agents via Observational Context Compression

m-a-p

Multimodal Art Projection

Submitted by

1f

WavAlign: Enhancing Intelligence and Expressiveness in Spoken Dialogue Models via Adaptive Hybrid Post-Training

·
12 authors

1

Submitted by

taesiri

SWE-chat: Coding Agent Interactions From Real Users in the Wild

·
6 authors

Submitted by

taesiri

Cortex 2.0: Grounding World Models in Real-World Industrial Deployment

·
28 authors

Submitted by

xiachongfeng

SAVOIR: Learning Social Savoir-Faire via Shapley-based Reward Attribution

hkuhk

The University of Hong Kong

Submitted by

deqing

Convergent Evolution: How Different Language Models Learn Similar Number Representations

UniversityofSouthernCalifornia

University of Southern California

Submitted by

FaisaI

Tadabur: A Large-Scale Quran Audio Dataset

·
1 authors

Submitted by

taited

ReImagine: Rethinking Controllable High-Quality Human Video Generation via Image-First Synthesis

GAP-LAB

Submitted by

dddraxxx

Visual Reasoning through Tool-supervised Reinforcement Learning

amazon-agi

1

Submitted by

kjappelbaum

AI scientists produce results without reasoning scientifically

jablonkagroup

Lab of Kevin Jablonka at Uni Jena

1

Submitted by

nielsr

Scaling Test-Time Compute for Agentic Coding

Submitted by

ayyyq

Self-Evolving LLM Memory Extraction Across Heterogeneous Tasks

UniversityofSouthernCalifornia

University of Southern California

Submitted by

taesiri

MMCORE: MultiModal COnnection with Representation Aligned Latent Embeddings

·
11 authors

Submitted by

yujiazheng

Diverse Dictionary Learning

CarnegieMellonU

Carnegie Mellon University

Submitted by

zhaihaotian

Abstain-R1: Calibrated Abstention and Post-Refusal Clarification via Verifiable RL

minnesotanlp

Submitted by

akhaliq

Image Generators are Generalist Vision Learners

·
25 authors

Submitted by

taesiri

CreativeGame:Toward Mechanic-Aware Creative Game Generation

·
9 authors