関係筋によると、AI企業Anthropicは2週間以内に新たな資金調達ラウンドを実施する可能性があり、評価額は9000億ドルを超える見通しだ。これが実現すれば、AI業界における同社の急成長を改めて示すことになる。
According to sources, Anthropic could close a new funding round within two weeks at a valuation exceeding $900 billion, underscoring the AI company's rapid ascent in the industry.
AI要約 ChatGPTが回答を拒否した質問に対し、AIエージェントが自ら課金して別サービスを利用し回答を得る仕組みを試した記事。MCPを活用してエージェントが自律的に支払いを行い、制限を回避するアプローチを紹介している。
EN The author built an AI agent that autonomously pays for external services via MCP to answer questions ChatGPT refused to handle, demonstrating how agents can bypass restrictions through self-initiated payments.
og AI要約 ThunderbirdをMCPサーバー化し、Claude Desktopからメールの検索・閲覧・送信などを操作できるようにする手法を紹介する記事。MCP連携の設定手順や活用例を解説している。
EN This article introduces how to turn Thunderbird into an MCP server so that Claude Desktop can operate emails such as searching, reading, and sending. It explains the setup procedure and usage examples for the MCP integration.
og AI要約 物理情報ニューラルネットワーク(PINN)のメタ学習において、タスク間の異質性が学習を阻害する問題に対し、合成的なメタ学習手法を提案。共有成分とタスク固有成分に分離することで、多様な物理タスクへの汎化性能を向上させる。
EN This paper proposes a compositional meta-learning approach for Physics-Informed Neural Networks (PINNs) that mitigates task heterogeneity by decomposing knowledge into shared and task-specific components, improving generalization across diverse physics tasks.
og AI要約 本論文ではバイナリスパイキングニューラルネットワーク(SNN)を因果モデルとして定式化する新たな枠組みを提案。スパイク発火を介入として扱うことで、SNNの動作を因果推論の観点から解釈・解析できることを示している。
EN This paper proposes a novel framework that formulates binary spiking neural networks (SNNs) as causal models, treating spike events as interventions to enable causal-inference-based interpretation and analysis of SNN behavior.
og AI要約 実際の光学実験装置を用いて、仮説生成から実験実行までを自律的に行うエンドツーエンドの科学発見システムを提案。AIが物理的な計測機器を直接操作し、人間の介入なしに新しい科学的知見を導く可能性を示した研究。
EN This paper presents an end-to-end autonomous scientific discovery system operating on a real optical platform, where AI directly controls physical instruments to generate hypotheses and conduct experiments without human intervention.
og AI要約 思考から実行までを担う自律的な機械学習パイプライン生成手法を提案。複数のAIエージェントが協調し、エラー検出と自己修復機能を備えることで、人手を介さずにロバストなMLワークフローを構築できる枠組みを示している。
EN This paper proposes an autonomous ML pipeline generation framework using a self-healing multi-agent AI system, where collaborating agents detect and recover from errors to build robust end-to-end machine learning workflows without human intervention.
og AI要約 分散型AIサービスを実現するためのフレームワーク「TRUST」の初期バージョンを提案する論文。信頼性・透明性・セキュリティを備えたAIサービス基盤の設計指針を示している。
EN This paper introduces TRUST v0.1, a framework for decentralized AI services that aims to provide trust, transparency and security as foundational design principles for distributed AI infrastructure.
og AI要約 プログラミング学習中の学生がAIに助けを求める過程を分析した研究。学生とAIの対話における「Vibe Coding」現象を調査し、助言要請の行動パターンや学習への影響を明らかにする。
EN This study analyzes help-seeking processes in student-AI interactions during programming, examining the phenomenon of 'Vibe Coding' and its patterns and implications for learning.
AI要約 本研究は、自律取引エージェント群に対するストップロスおよび利確のパラメータを最適化する手法を提案する。多数のエージェントが協調して取引する環境下で、リスク管理パラメータを最適化することで、収益性とリスク制御の両立を図る。
EN This paper proposes a method for optimizing stop-loss and take-profit parameters for swarms of autonomous trading agents, aiming to balance profitability and risk control in cooperative multi-agent trading environments.
EN arXiv:2604.27151v1 Announce Type: new Abstract: Computer-use agents provide a promising path toward general software automation because they can interact directly with arbitrary graphical user interfa
EN arXiv:2604.27156v1 Announce Type: new Abstract: Rational belief revision is commonly viewed as being based on a preference order between possible worlds, with the resulting new belief set being those
EN arXiv:2604.27195v1 Announce Type: new Abstract: Accurate prediction of conversion from Mild Cognitive Impairment (MCI) to Alzheimers Diseases (AD) is essential for early intervention, however, develop
EN arXiv:2604.27217v1 Announce Type: new Abstract: Cognitive decline is highly heterogeneous across individuals, which complicates prognosis, trial design, and treatment planning. We present the Personal
EN arXiv:2604.27221v1 Announce Type: new Abstract: Agentic web search increasingly faces two distinct demands: deep reasoning over a single target, and structured aggregation across many entities and het
EN arXiv:2604.27228v1 Announce Type: new Abstract: Democratic discourse analysis systems increasingly rely on multi-agent LLM pipelines in which distinct evaluator models are assigned adversarial roles t
AI要約 本番システムで使用中のLLMが提供終了(EOL)を迎えた際の、新モデルへの移行を支援するフレームワークを提案。性能維持を保証しながら信頼性のあるモデル切替を実現する手法を示している。
EN This paper proposes a framework for confidently migrating LLMs in production systems when models reach end-of-life, ensuring performance preservation and reliable transitions between models.
og EN arXiv:2604.27126v1 Announce Type: cross Abstract: This study presents an unsupervised machine learning workflow for electrofacies analysis in the offshore Keta Basin, Ghana, where core data are scarce
EN arXiv:2604.26984v1 Announce Type: new Abstract: Representational collapse, where embeddings become anisotropic and lose multi-scale structure, can erode downstream performance long before performance
EN arXiv:2604.26985v1 Announce Type: new Abstract: Masked diffusion models (MDMs) generate discrete sequences by iterative denoising under an absorbing masking process. In standard masked diffusion, if a
EN arXiv:2604.26991v1 Announce Type: new Abstract: Recent advances in data-centric medical AI have produced highly accurate diagnostic systems, but the emphasis on data curation and performance metrics h
EN arXiv:2604.27003v1 Announce Type: new Abstract: Memory-augmented LLM agents offer an appealing shortcut to continual learning: rather than updating model parameters, they accumulate experience in exte
EN arXiv:2604.27011v1 Announce Type: new Abstract: AutoML, intended as the process of automating the application of machine learning to real-world problems, is a key step for AI popularisation. Most Auto
EN arXiv:2604.27014v1 Announce Type: new Abstract: The scarcity of high-quality annotated medical data, particularly in mental health, poses a significant bottleneck for training robust machine learning
EN arXiv:2604.27019v1 Announce Type: new Abstract: Safety-aligned language models must refuse harmful requests without collapsing into broad over-refusal, but the training-time mechanisms behind this tra
EN arXiv:2604.27031v1 Announce Type: new Abstract: In a continual learning setting, we require a model to be plastic enough to learn a new task and stable enough to not disturb previously learned capabil
EN arXiv:2604.27033v1 Announce Type: new Abstract: Deep learning for cross-subject EEG decoding is hindered by high inter-subject variability, which introduces a severe domain shift between training and
EN arXiv:2604.27045v1 Announce Type: new Abstract: As Large Language Model (LLM) agents transition from single-session tools to persistent systems managing longitudinal healthcare journeys, their memory
EN arXiv:2604.27063v1 Announce Type: new Abstract: Continual learning agents with finite capacity must balance acquiring new knowledge with retaining the old. This requires controlled forgetting of knowl
EN arXiv:2604.27077v1 Announce Type: new Abstract: The Normalized Transformer, or nGPT (arXiv:2410.01131) achieves impressive training speedups and does not require weight decay or learning rate warmup.
EN arXiv:2604.27083v1 Announce Type: new Abstract: RLVR and OPD have become standard paradigms for post-training. We provide a unified analysis of these two paradigms in consolidating multiple expert cap