The Research Desk.

The most upvoted and starred AI research crossing the community today.

Last Brew Time: Aug 8, 2026, 10:25 AM PT

AlphaXiv Trending

Leanstral
AlphaXiv
77

Leanstral

Aditi Kabra, Albert Q. Jiang, Andrew Zhao

Machine Learning
Towards Physics of Multimodal Pretraining: Knowledge Flow, Modality Synergy, Early Unification, and Recipes
AlphaXiv
72

Towards Physics of Multimodal Pretraining: Knowledge Flow, Modality Synergy, Early Unification, and Recipes

Junlin Han, Shengbang Tong, David Fan

FAIR, Meta, Reality Labs, Meta, University of Oxford

Computer Vision
On-Policy Self-Distillation without Any Supervision
AlphaXiv
37

On-Policy Self-Distillation without Any Supervision

Yijiang Li, Bingyang Wang, Yijun Liang

Georgia Institute of Technology, University of Maryland, College Park, ByteDance

NLP
Test-Time Scaling in Reasoning LLMs: Inference Regimes, Evaluation, and Reproducibility
AlphaXiv
37

Test-Time Scaling in Reasoning LLMs: Inference Regimes, Evaluation, and Reproducibility

Mohsen Hariri, Weicong Chen, Nahal Shahini

Case Western Reserve University

DreamWAM: Beyond RGB Future Prediction for World Action Models
AlphaXiv
36

DreamWAM: Beyond RGB Future Prediction for World Action Models

Shanglin Yuan, Weiheng Zhao, Xin Shi

Computer Vision
OPD-V: Visual On-Policy Self-Distillation with Modality Balance
AlphaXiv
32

OPD-V: Visual On-Policy Self-Distillation with Modality Balance

Aniri, Jinhe Bi, Peng Liao

National University of Singapore, Ludwig Maximilian University of Munich, Munich Center for Machine Learning, Sun Yat-sen University

HuggingFace Daily Papers

Reinforcement Learning
DataSpace: Benchmarking Data Agents for Verifiable Analytics over Heterogeneous Workspaces
HuggingFace
24

DataSpace: Benchmarking Data Agents for Verifiable Analytics over Heterogeneous Workspaces

Boyan Li, Zhuowen Liang, Yupeng Xie, Xiaotian Lin, Tianqi Luo

HKUSTDial

Reinforcement Learning
Activity Frames: Deterministic Screen-Activity Compilation for Agent Memory and Replay
HuggingFace
16

Activity Frames: Deterministic Screen-Activity Compilation for Agent Memory and Replay

Nossa Iyamu

NLP
KVAE: Family of Tokenizers for Multimodal Generative Models
HuggingFace
14

KVAE: Family of Tokenizers for Multimodal Generative Models

Andrey Shutkin, Denis Parkhomenko, Ivan Kirillov, Kirill Chernyshev, Kirill Malakhov

Kandinsky Lab

Reinforcement Learning
FactorJEPA: Factorizing Monolithic Futures into Layout-Agent-Interaction Channels for Crowded and Chaotic Global South Urban Worlds
HuggingFace
5

FactorJEPA: Factorizing Monolithic Futures into Layout-Agent-Interaction Channels for Crowded and Chaotic Global South Urban Worlds

Kapil Wanaskar, Gaytri Jena, Aman Chadha, Vinija Jain, Vasu Sharma

Robotics
Weights or Skills? A Survey of Robot-Learning Techniques: from Action-Predicting Weights to Robots that Write their Own Skills
HuggingFace
4

Weights or Skills? A Survey of Robot-Learning Techniques: from Action-Predicting Weights to Robots that Write their Own Skills

Gaytri Jena, Kapil Wanaskar, Vinija Jain, Aman Chadha, Vasu Sharma

Computer Vision
GaussianSelector: Lightweight Human-Guided Object Selection in 3D Gaussian Splatting with Graph Optimization
HuggingFace
4

GaussianSelector: Lightweight Human-Guided Object Selection in 3D Gaussian Splatting with Graph Optimization

Baihan Yang, Tiexin Li, Yuheng Liu, Xin Lin, Xinke Li

LL Lab

AI Research Papers — August 8, 2026 Edition | Agentic Brew