新モデル・リリース
A
291 件中 271〜291 件目を表示
-
Evaluating DeepSeek V4.1 Flash-1017 for loss of control and harmful manipulation risks
-
AWSとGoogle Cloudが「PyTorch」のAWS Trainium、Google Cloud TPU対応版をオープンソースで公開。NVIDIAからの移行に挑むAWSとGoogle CloudがPyTorchの自社AIチップ対応版をOSS公開10月20日・21日に米カリフォルニア州で開かれたPyTorch Conference 2026で、AWSとGoogle Cloudが相次いで、PyTorchの自社AIプロセッサ向け対応版をオープンソースとして発表した。AWSはTrainium、Google CloudはTPUに対応し、NVIDIAからの移行を狙う動きとされる。
-
TRON: Transferring Naturally Evolved Network Architectures into Neuroevolution
-
Adapt for Planning, Not just Prediction: Optimistic Test-Time World Model Adaptation
-
Offline Language-Based Support System for Tactical Decision-Making in Air Combat Operations
-
Hierarchical Skill Evolution for Coding Agent
-
Making Reasoning Visible: Configuration-Aware Diagnostic Modeling for Adaptive Scaffolding in Structured Reasoning Tasks
-
ArchiveLens: A Generative Spatial Knowledge Framework for Semantic Reconstruction of Historical Maps
-
SAP-ViT: A Self-Attributing Patch Vision Transformer for Medical Image Classification
-
ExCal: Optimization-Free Multivariate Time Series Anomaly Detection via Excess Calibration
-
Taking Japan-Born AI Products to the World: An Interview with Sakana AI’s VP of GTMSakana AI、GTM責任者に聞く研究発の製品を世界へ届ける戦略Sakana AIは2026年にSakana Chat、Marlin、Fugu、Translateを相次いで発表し、研究開発企業から企業・官公庁向けに製品を届ける段階に入った。9月に本格始動したGTMチームを率いるVP柿島氏は、AWS Japan、SORACOM、Skydioで新技術の事業化を担ってきた人物。インタビューでは、なぜ今Sakana AIなのか、研究由来の製品をどう売るか、仕組みが未整備のチームで働く面白さを語る。
-
LoopFold: Parallelizing Recurrent Depth in Looped Transformers
-
Cross-Domain Sequential Recommendation with Representation-Level User Transfer
-
StateLoop: Persistent State for Looped Transformers
-
Beyond Modeling Order: Dual-Path Spatiotemporal Representation Learning for Cross-Subject EEG Emotion Recognition
-
CANS: Capability-Preserving Neuron Steering for Large Language Models
-
S3AC: Stabilized, Selective, and Scalable Actor-Critic for LLM Reasoning
-
MATCHA: A Chinese Multi-Disorder Mental Health Dataset with Cross-Lingual Analyses on Social Media
-
Beyond Words Alone: Disentangling Distributed Evidence for Visually Grounded Cultural Translation
-
ImgMT: Wikipedia-based Dataset For Evaluating Text-in-Image Translation
-
The Geometry of Emotion Representations Across Scale, Depth, and Instruction Tuning