AI Wiki — Knowledge Base

創業以来、オープンソースコミュニティから多くを学び、また研究成果の公開を通じてそこに貢献してきました。オープンなエコシステムが健全なAI産業と技術主権を支える重要な基盤の一つであると考えており、その発展を支持します。このたび、Sakana AIは、オープンウェイトAIモデルに関…

創業以来、オープンソースコミュニティから多くを学び、また研究成果の公開を通じてそこに貢献してきました。オープンなエコシステムが健全なAI産業と技術主権を支える重要な基盤の一つであると考えており、その発展を支持します。このたび、Sakana AIは、オープンウェイトAIモデルに関する公開書簡「Open Weights and American AI Leadership」に署名しました。書簡は

safetydavid-ha--x

27 Jul 2026

Search: “models”

Sling2Sim2Real: One-Shot Elastic System Identification for Non-Destructive Slingshot Policy Learning

SMART: LLM-Augmented Hybrid Retrieval for Dynamic Product Ads

Smooth Learning with Hard Constraints via Legendre-Regularized Policies

Spatio-Temporal Conditional Denoising Transformer for Modality-Missing RGBT Tracking

spec: add DSpark speculative decoding by wjinxu · Pull Request #25173 · ggml-org/llama.cpp

SpecFormer: Mitigating Embedding and Attention Collapse via Spectral-Aware Transformer for Recommendation

Speech Signals Complement LLMs for Predicting Interpersonal Attraction in Speed Dating

SPRKD: Effective Knowledge Distillation for Deep Neural Networks via Saddle Region Approximation

Structural Preservation Governs Data Augmentation in Deep Learning-Based Laser Speckle Material Classification

Subject-Level Heterogeneity in EEG Motor Imagery Decoding: A Large-Scale Benchmark and Portfolio-Based Reduction of the Search Space

SymStep: Symbolic Step Verification for Logical Reasoning

TableMind: An Autonomous Programmatic Agent for Tool-Augmented Table Reasoning

The administration is very consistent: All aspects of foreign involvement in AI are banned: people (students and employees), hardware and mo…

The Intruder Threshold: A Spectral Law for LoRA Fine-Tuning

The Label Complexity of Class-Conditional Coverage under Distribution Shift

The OlmoEarth Platform: Geospatial inference at planetary scale

ThinkingCap-Qwen3.6-27B warrants a look

TimePLE: Rethinking Temporal Representation for Video Temporal Grounding

Tines seeks to tame ‘wild code’ AI sprawl

TRACE: Business Rule-Grounded Reasoning Curriculum for Knowledge-Preserving Parametric Tool Retrieval in Enterprise LLMs

Two-Timescale Hierarchical Reinforcement Learning for Resilient Operations

Unsupervised Graph Representation Learning with Complementary View Alignment

UP-Fuse: Uncertainty-guided LiDAR-Camera Fusion for 3D Panoptic Segmentation

Variational Quantum Conditional Boltzmann Machines for Time-Series Forecasting: Architectures, Symmetric Hyperparameter Evaluation, and a Nonlinear Benchmark

When Can Depth Replace Precision? A Resource Theory of Quantized Neural Computation

Who Pays the Price? Stakeholder-Centric Prompt Injection Benchmarking for Real-world Web Agents

X-Stage: An Overlooked Pipeline Stage for Communication-Computation Overlap in DiT Inference

Zuck's opinion: The AI Future Is for Everyone

A Framework for Individual Tree Growth Reconstruction Using Multi-Platform Laser Scanning

A Graph-Based Control Interface for Traffic Signals on Heterogeneous Road Networks

Adaptive Contrast Enhancement and Optimised Feature Matching for RootSIFT-Based Palm-Vein Recognition

Adversarial Prompts for Acceptance Collapse in Speculative Decoding

AgentHOI: Multi-Agent Reasoning for Human-Object-Interaction Video Generation via Implicit Representation Alignment

All of this took me under 5 minutes to find. Great reminder to all of us to be more vigilant of multiplayer mode in AI and to maybe ask Code…

An Introduction to Bayesian and Frequentist Simulation-Based Inference with Machine Learning

At a small business, the person closest to a problem is often the one who has to solve it—even when it falls outside their job description. …

b10145

b10146

b10149

b10150

b10151

b10152

b10154

Bilateral Trade Under Heavy-Tailed Valuations: Minimax Regret with Infinite Variance

Claude chats showed in Google and Bing search results, despite Anthropic's robots.txt saying not to crawl them, likely because the pages lacked a 'noindex' tag (Maddy Varner/Wired)

Closing the Loop: Training-Free Revisit Consistency for Autoregressive Generative Rendering

DeepUrban: Interaction-Aware Trajectory Prediction and Planning for Automated Driving by Aerial Imagery