The Research Desk.

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

Last Brew Time: Oct 28, 2026, 12:16 PM PT

X.com Research Buzz

NLP
Why are all LLMs Obsessed with Japanese Culture? On the Hidden Cultural and Regional Biases of LLMs
X.com
12696

Why are all LLMs Obsessed with Japanese Culture? On the Hidden Cultural and Regional Biases of LLMs

Evidence for the Representation of Non-Hierarchical Structures in Language
X.com
7568

Evidence for the Representation of Non-Hierarchical Structures in Language

Morten H. Christiansen, Yngwie A. Nielsen

Cornell University, Aarhus University

NLP
STAIR (STructure Aware Information Retriever): A novel dataset and LLM based retriever for document structure augmentation
X.com
6703

STAIR (STructure Aware Information Retriever): A novel dataset and LLM based retriever for document structure augmentation

IBM

AlphaXiv Trending

Reinforcement Learning
Work Agents: Systems, Capabilities, and Optimization
AlphaXiv
47

Work Agents: Systems, Capabilities, and Optimization

Views

Robotics
Learning Force-Reactive Humanoid Loco-Manipulation
AlphaXiv
34

Learning Force-Reactive Humanoid Loco-Manipulation

Views

Affiliation: Amazon FAR (Frontier AI & Robotics) New York University Carnegie Mellon University University of Chicago, Amazon FAR team co-lead, Amazon FAR (Frontier AI & Robotics) New York University Carnegie Mellon University University of Chicago

Reinforcement Learning
Evolving Agent Teams
AlphaXiv
18

Evolving Agent Teams

Views

Affiliation: University of California, Berkeley, University of California, Berkeley

Reasoning
ThinkWorld: World Imagination Reinforces Chain-of-Dynamics Reasoning
AlphaXiv
14

ThinkWorld: World Imagination Reinforces Chain-of-Dynamics Reasoning

Views

Affiliation: NVIDIA, Affiliation: NVIDIA Affiliation: KAIST, NVIDIA

Timestral
AlphaXiv
9

Timestral

Computer Vision
3D Skew-Normal Splatting
AlphaXiv
6

3D Skew-Normal Splatting

Affiliation: Fudan University, Fudan University

HuggingFace Daily Papers

Retrieval
EmbeddingGemma 2: Unified and Efficient Multimodal Representations
HuggingFace
10

EmbeddingGemma 2: Unified and Efficient Multimodal Representations

Henrique Schechter Vera, Sahil Dua, Jovan Andonov, Sara Smoot, Sebastian Russo

Google

Machine Learning
Beyond the Estimate: Wiener-Informed Decoding for Easier Training of Pixel-Space Flow Models
HuggingFace
8

Beyond the Estimate: Wiener-Informed Decoding for Easier Training of Pixel-Space Flow Models

Ziqiao Wang, Yang You

National University of Singapore

Reinforcement Learning
What Outcome RL Buys: Finishing, Reliability, Rarely Reach
HuggingFace
2

What Outcome RL Buys: Finishing, Reliability, Rarely Reach

Shan Yang

Efficiency
How Much Does the Hint Matter? Scope Conditions for Hindsight Self-Distillation in Financial Reasoning
HuggingFace
2

How Much Does the Hint Matter? Scope Conditions for Hindsight Self-Distillation in Financial Reasoning

Johnny Yu, Cyril Gorlla

CTGT, Inc.

WSU Carbon Lab
Sub-monolayer tracking of molecular states at buried interfaces
HuggingFace
1

Sub-monolayer tracking of molecular states at buried interfaces

Harlan Heilman, Em Biegler, Sushil Adhikari, Victor Murcia, Tanner Melody

WSU Carbon Lab

Speech Audio
Lipi-Ghor: A Large-Scale, Diarized, Multi-Domain Bengali Speech Corpus and a Scalable Pipeline for Building It
HuggingFace
0

Lipi-Ghor: A Large-Scale, Diarized, Multi-Domain Bengali Speech Corpus and a Scalable Pipeline for Building It

Sanjid Hasan, A H M Fuad, Risalat Labib, Bayazid Hasan Simon

Team Villagers