AI Wiki — Knowledge Base

チームのLLM開発が本格化してきたので仲間を募集します！人数を必要以上に増やさず、一人ひとりが大きな役割を担うチームを作っています。高い基礎力と知的好奇心、そして違った強みを持つメンバーが集まって、全体の目的に向かって高い主体性を持って動いています。そんなチームの一員になれます…

チームのLLM開発が本格化してきたので仲間を募集します！人数を必要以上に増やさず、一人ひとりが大きな役割を担うチームを作っています。高い基礎力と知的好奇心、そして違った強みを持つメンバーが集まって、全体の目的に向かって高い主体性を持って動いています。そんなチームの一員になれます。僕自身もこのチームでメンバーと仕事することが何より楽しいです。 LLM経験者はもちろん歓迎。未経験でも、しばらくLL

researchdavid-ha--x

22 Aug 2026

research

Rethinking Expressivity and Efficiency in Test-Time Training

Robust Discovery of Coarse-Grained Continuum Equations from Microscopic Dynamics

SAC-Copula: Quality-Preserving Watermarking for Diffusion Language Models via Smooth Correlated Gumbel Fields

SAGE: A Unified Algebra and Self-Adaptive Execution for AI Functions in SQL

Scaling Muon for Diffusion Transformers

SCOPE: A Generative Approach for LLM Prompt Compression

Self-Supervised Speech Representations Track Spoken Language Convergence to Adult Models in Infants and Children Who Are Deaf/Hard-of-Hearing

Shared Physics Responses Recover Hidden Rankings in Neural Operator Libraries

Sharing the Control Authority Between Deep Reinforcement Learning and Model Predictive Control: Application to Multi-Class Transportation Networks

Shortcut Learning in a Public Grape Disease Dataset: Annotation Granularity as a Modulator, Not a Cause

Significant Other AI: Identity, Memory, and Emotional Regulation as Long-Term Relational Intelligence

SKILL-RAG: Self-Knowledge Induced Learning and Filtering for Retrieval-Augmented Generation

SPARC: Single-Pass Scaling for Motion Forecasting with Conformal Bayesian Last Layers

SPARCL: Spectral Partitioned Analytic Continual Learning

Sparse Light Field Sampling Improves Casual 3D and 4D Reconstruction

Spatially Aware Dictionary-Free Koopman Eigenfunction Identification for Modeling and Control

SPD Matrix Learning for Neuroimaging Analysis: Perspectives, Methods, and Challenges

SuppreSensing: Expert-Guided Feature Recalibration and Discrepancy Augmentation for Multimodal Object Detection

Temporal Validity on Real Software Histories: Eliminating Stale-Fact Errors in Code-Assistant Memory over GitHub Fixes

The Cost of a Physics Prior Is Bounded by the Ablation Gap

The Download: kids outlearning AI, and space travel agents

Time-Aware Tranformer-Based Prediction Model for AECOPD

Toward Auto-Research: Mining Falsifiable Research Ideas from Paper Knowledge Graphs with Categorical Structure

Towards Investigating Residual Hearing Loss: Quantification of Fibrosis in a Novel Cochlear OCT Dataset

TRACE-C: Rank-Calibrated Relational Anomaly Detection for Multi-Stream Operational Telemetry

TracingFlow: A Simulation-Free Trajectory Inference Framework Based on Second-Order Dynamics

Training DeepFilterNet with Accurate Room Acoustic Simulations Improves Single-Channel Speech Enhancement

TreeWY: Speculative Verification for Gated DeltaNet Hybrids

Trilingual Topic Modeling of Sri Lankan Parliamentary Debates

TriPLU: Bypassing the Gate with Direct Trilinear Product FFNs in Tiny Language Models

Truthful Calibration Measures for Sequential Prediction

Tydra: An Efficient Hybrid Model for Tabular Data

Using Human-LLM Disagreement to Improve Checklist-Based Quality Appraisal

Vision Foundation Model Driven Foreground-Aware Pseudo-LiDAR Generation for Monocular 3D Object Detection

When Clean Data Hurts: Learning with Monotone Corruptions Beyond Binary Classification

When Graph-JEPA Learns the Wrong Thing: Diagnosing and Repairing Category-Conditional Collapse

When Looks Do Not Lie: Discourse Structure Guided In-Context Learning for Faithful Diagram Generation

When the Feature Pool Goes Algorithmic: Extending Mufwene's Ecology of Language Evolution to LLM-Mediated Exposure

When to Ponder: Adaptive Compute Allocation for Code Generation via Test-Time Training

When Trust Meets Truth: Trust-Truth Separability in LLM-as-Judge

Implementing Watermarking for Language Models [P]

More data than open-source AI is taking share from OpenAI and Anthropic. Open source has gone from 28% token share to 62% token share @verce…

Scientific terms should have precision. If we use the terms VLM, VLA, WAM in an indiscriminate fashion, as is becoming common in robotics, w…

You don't often see one-word titles in AI papers. That aside, strong recommend this paper from Google DeepMind. I think this is an interesti…

We’re Hiring: Member of Technical Staff (LLM Development) 🐟 https://sakana.ai/careers/member-of-technical-staff-llm-development/#English Th…

4DAnyone: Create Anyone in 4D from a Casual Monocular Video

A Plug-in Interpretation of Conditioning in Score-Based Diffusion Models