AI Wiki — Knowledge Base

Harness EngineeringからRSIへ AI自身がAIをつくり、改善していく──再帰的自己改善（RSI）はどのように実現するのでしょうか。この問いに対して、Thinking Machines Lab 共同創業者の Lilian Weng 氏（@lilianweng）が…

Harness EngineeringからRSIへ AI自身がAIをつくり、改善していく──再帰的自己改善（RSI）はどのように実現するのでしょうか。この問いに対して、Thinking Machines Lab 共同創業者の Lilian Weng 氏（@lilianweng）が、モデルを取り巻く実行系＝「ハーネス」の設計こそが現実的な出発点だと論じるブログ記事を公開しました。RSI周辺の研究が

researchdavid-ha--x

7 Jul 2026

Search: “research”

Exploring Context-aware and LLM-driven Locomotion for Immersive Virtual Reality

Exploring the Rashomon Set for Concept-Based Models

Fast, Parallel, Query-Efficient Binary Classification

FastCSP: Accelerated Molecular Crystal Structure Prediction with Universal Model for Atoms

Feature weighting for data analysis via evolutionary simulation

FedFFT: Taming Client Drift in Federated SAM via Spectral Perturbation Filtering

Feeling the Unexpected: ResTacVLA for Contact-Rich Manipulation via Residual Tactile Representation

Fields of the Planet: Field Boundary Mapping Beyond 10m

Flash-BoN: Instant Drafts for Inference-Time Scaling in Diffusion Models

FLASH: Flexible Learning of Adaptive Sampling from History in Temporal Graph Neural Networks

FlashBlock: Attention Caching for Efficient Long-Context Block Diffusion

Folding, Reasoning, and Scaling with Open-source Drug Discovery Engine

Foundations of Equivariant Deep Learning: Unifying Graph and Sheaf Neural Networks

Fourier Splatting: Generalized Fourier encoded primitives for scalable radiance fields

Framework of Thoughts: A Foundation Framework for Dynamic and Optimized Reasoning based on Chains, Trees, and Graphs

From Gentlemen to Frontiermen: Masculine Formations in English-Language Fiction (1771--1930)

From Interaction to Intent: Inferring User Objectives from Provenance Logs

From Mobile Data to Business Insights: An End-to-End Analytics Framework for Large-Scale Urban Mobility Analysis and Decision Support

From Tensor Buffer to Distributed Memory Hierarchy: A Survey of KV Cache Management for LLM Serving

Full Glyph Images Beat Token Embeddings: A Controlled Study for Transformers

Function-Space Diffusion for Motion Planning

Fusion: A Framework for Unified Sequential Token AdaptatIon in VisiOn TraNsformers

GDBR: Label Recovery Attack Against Partial Gradient Encryption in Federated Learning

GeoFlow: Geo-Aware Modeling of Inter-Area Relationships in Origin-Destination Flow Prediction and Generation

Geographic Diversity Beats Data Volume for Cross-Domain Generalization in Zero-Label JEPA Driving World Models

Geometric Reciprocity: Unlocking Self-Supervision for Stereoscopic Video Generation

Geometry-Aware Bayesian Quantification via Compositional Data Analysis

Geometry-Aware Visual Odometry for Bronchoscopic Navigation via High-Gain Observer Fusion

GeoSAM-Lite: A Lightweight Foundation Model for Onboard Remote Sensing Segmentation

GestaltMML: Enhancing Rare Genetic Disease Diagnosis through Multimodal Machine Learning Combining Facial Images and Clinical Text

GlacierCastAI: Predicting Glacier Retreat from Multi-Modal Satellite Imagery and Climate Signals

GlaKG: A Biomarker-Centric Fundus Knowledge Graph for Explainable Glaucoma Diagnosis and Risk Assessment

Glance-Say: Multimodal Human-Robot Collaboration and Intent Recognition via Sticky Glance

Global Pose Control for Generative View Synthesis in Normalized Object Coordinate Space

Global universal approximation with Brownian signatures

Gradient Descent as Implicit EM in Distance-Based Neural Models

Granger Causality in Extremes

Graph Unitary Message Passing

GRASP: Graph-Reasoning Aided Survey Planning for High-Fidelity Related Work Generation

Green for Go, Red for No: Visual Grounding via Semantic Segmentation for VLA Navigation Policies

Greenland isn’t surrounded by Chinese and Russian ships. There hasn’t been $19.2 trillion invested in the US under Trump. Trump hasn’t settl…

GUSH3R: Everyone Everywhere All at Once as Gaussians

Handwriting Trajectory Recovery with Diffusion Models

Harness EngineeringからRSIへ AI自身がAIをつくり、改善していく──再帰的自己改善（RSI）はどのように実現するのでしょうか。この問いに対して、Thinking Machines Lab 共同創業者の Lilian Weng 氏（@lilianweng）が…

HASSL: Hierarchy-Aware Self-Supervised Learning Framework for Single Cell Microscopy

Heterogeneous Graph Condensation via Role-Aware Clustering

Hierarchical Evidence-Driven Reasoning for Long Document Understanding

Hierarchical Sparse Attention Done Right: Toward Infinite Context Modeling