#1
Gemini 3 Deep Think: Advancing science, research and engineering
<img src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/gemini-3_deep-think_keyword_hea.max-600x600.format-webp.webp">We’re releasin...
2026年7月4日 本日の厳選記事 20件
🌏 International AI News (English)
#2
New research shows how AMIE, our medical AI, could help manage health conditions.
<img src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/AMIE_Mx_Nature_Social_Visual_Va.max-600x600.format-webp.webp">Research in “N...
#3
Beyond the Library: An Agentic Framework for Autoformalizing Research Mathematics
arXiv:2606.31134v1 Announce Type: new Abstract: While Large Language Models (LLMs) have demonstrated exceptional capabilities in mathematical reasonin...
#4
New ways to balance cost and reliability in the Gemini API
<img src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/cost_reliability_Gemini_API-soc.max-600x600.format-webp.webp">Google is intr...
#5
How animators and AI researchers made ‘Dear Upstairs Neighbors’
<img src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/DUN_poster_16x9_v02.max-600x600.format-webp.webp">Today, our animated short ...
#6
Reduce friction and latency for long-running jobs with Webhooks in Gemini API
<img src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/WebhooksGeminiAPI-hero.max-600x600.format-webp.webp">Event-Driven Webhooks a...
#7
We're launching two specialized TPUs for the agentic era.
<img src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/Were_launching_two_TPUs_social.max-600x600.format-webp.webp">The eighth gene...
#8
100 things we announced at I/O 2026
<img src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/100_things_Social.max-600x600.format-webp.webp">This year at Google I/O 2026...
#9
The latest AI news we announced in January
<img src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/LatestAI_v5.max-600x600.format-webp.webp">Google AI announcements from Janua...
#10
The latest AI news we announced in June 2026
<img src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/260701_ICYMI-June-AI_Thumb.max-600x600.format-webp.webp">Here are Google’s l...
🇯🇵 日本語AIニュース
#1
OpenAI、科学論文執筆環境「Prism」公開 「GPT-5.2」搭載のLaTeX対応ワークスペース
OpenAIは、「GPT-5.2」を統合した科学者向けのAI搭載LaTeXエディタ「Prism」を無料公開した。論文の構造や数式、文献情報をAIが把握し、執筆や推敲、手書き数式のコード変換などを1つのワークスペースで完結させる。個人ユーザーは無制限で利用可能だ。
#2
Google、「Gemini 3.1 Pro」搭載の自律型リサーチAI「Deep Research／Research Max」発表
Googleは「Gemini 3.1 Pro」ベースの自律型リサーチエージェント「Deep Research」と「Deep Research Max」を発表した。調査の自動化に特化し、社内データや専門的な金融データとの連携が可能。上位のMax版は高い推論能力を備え、複雑なレポート作成や高度なデータ視...
#3
#4
「Gemini 3.5」登場 軽量モデルは一部3.1 Pro超え、高性能モデルは6月公開予定
Googleが新しいAIモデル群「Gemini 3.5」シリーズを発表。軽量モデルの「Gemini 3.5 Flash」は発表同日から利用可能。高性能モデルの「Gemini 3.5 Pro」は6月にリリース予定。
#5
Anthropic、新AIモデル「Claude Sonnet 4.6」発表 “人間レベル”のPC操作も一部可能に
米Anthropicは、新たなAIモデル「Claude Sonnet 4.6」を発表した。前モデル「Claude Sonnet 4.5」に比べ、コーディングや自律的なPC操作などの性能が向上したという。
#6
ロボットが計器読み取り可能に、Googleが産業現場向けAIモデル「Gemini Robotics-ER 1.6」発表 Boston Dynamicsとも連携
Google DeepMindはロボット向けVLM「Gemini Robotics-ER 1.6」を発表。空間認識や計測器の読み取りなど、産業現場で求められる能力を高めたモデルだとうたう。
#7
Claude次世代モデル「Mythos」が一般公開されないワケ セキュリティ能力高すぎて「ゼロデイ攻撃自律開発」「出られないはずのサンドボックスから脱出」
Anthropicは、現行モデルを凌ぐ性能を持つ次世代モデル「Claude Mythos Preview」の存在を公表した。攻撃への悪用リスクから一般公開を見送り、現在は防御目的の限定活用にとどめている。
#8
Google、推論の深さを制御する「thinking levels」搭載の「Gemini 3.1 Flash-Lite」リリース
Googleは、Gemini 3シリーズ最速の「Gemini 3.1 Flash-Lite」をリリースした。「Gemini 2.5 Flash」と比較して出力が2.5倍高速化し、ベンチマークでもそれを上回る性能を達成。タスクに応じた推論の深さを制御できる「thinking levels」も搭載し、大...
#9
論文の図を生成するAI「PaperBanana」 Googleなどが開発 ブラウザ上で試せるジェネレーターも公開中
北京大学とGoogle Cloud AI Researchに所属する研究者らは、学術論文における図や統計プロットを自動生成するフレームワークを開発した研究報告を発表した。
#10
マイクロソフト、OpenAI の GPT-5 を一般ユーザー、開発者、エンタープライズ向け製品に統合
エリオット スミス (Elliott Smith) ※本ブログは、米国時間2025年8月7日（木）に公開された “Microsoft incorporates OpenAI’s GPT-5 into consumer, developer and enterprise offerings ̵...
最近の記事（履歴）
生成AI導入の10大リスクと対策！推進担当者のための実践ガイド
経営層から「生成AIで全社DXを進めろ」と指&#...
生成AI導入の社内マニュアルの作り方！必須の10項目を解説
経営から「全社で生成AIを使え」と号令が...
ClaudeでBlenderやPhotoshopを直接制御 Anthropicがクリエイティブ向け新コネクタ8件公開
Anthropicは、ClaudeにAdobe PhotoshopやBlenderなどと連携する8種類のクリエイティブ系「コネクタ」を導入した。AIが反復作業やデータ同期を担うことで、クリエイターは手動の工程を省き、迅速なアイデア創出が可能になるとしている。
AIエージェントを活用する中堅・中小は何を使っている？ 先進企業が「手放さないツール」【調査】
ノークリサーチの調査で、AIエージェントを既に導入している中堅・中小企業が併用するツールが判明した。AIエージェント活用企業が併用するツールとは何か。
「OpenAIの独占権」手放したMicrosoft、AI競争に勝てるか？ 巨額投資マネーの行方は
Google、Microsoft、Meta、AmazonによるAI投資が止まらない。ビッグテックは巨額の投資を回収できるのか。特にMicrosoftに厳しい目が向けられている。
AIに丸投げ……で失敗 漫画「1週間後に生成AIで恥をかく新入社員」【残り1日】
とある企業に入社した新入社員「ニイジマ」が、AI活用の“地雷”を踏みまくっていく様子を通して、会社におけるAIの使い方、その初歩を紹介します。
AI Identity: Standards, Gaps, and Research Directions for AI Agents
arXiv:2604.23280v1 Announce Type: new Abstract: AI agents are now running real transactions, workflows, and sub-agent chains across organizational bou...
Vibe Medicine: Redefining Biomedical Research Through Human-AI Co-Work
arXiv:2604.23674v1 Announce Type: new Abstract: With the emergence of large language models (LLMs) and AI agent frameworks, the human-AI co-work parad...
CiteRadar: A Citation Intelligence Platform for Researcher Profiling and Geographic Visualization
arXiv:2604.25057v1 Announce Type: new Abstract: Understanding the geographic reach and community structure of one's scholarly citations is increasingl...
Zero Shot Coordination for Sparse Reward Tasks with Diverse Reward Shapings
arXiv:2604.25076v1 Announce Type: new Abstract: Many Multi-Agent Reinforcement Learning (MARL) agents fail to adapt properly to cooperating with agent...
DiRe-RAPIDS: Topology-faithful dimensionality reduction at scale
arXiv:2604.25209v1 Announce Type: new Abstract: Dimensionality reduction methods such as UMAP and t-SNE are central tools for visualising high-dimensi...
An Intelligent Fault Diagnosis Method for General Aviation Aircraft Based on Multi-Fidelity Digital Twin and FMEA Knowledge Enhancement
arXiv:2604.22777v1 Announce Type: new Abstract: Fault diagnosis of general aviation aircraft faces challenges including scarce real fault data, divers...
PExA: Parallel Exploration Agent for Complex Text-to-SQL
arXiv:2604.22934v1 Announce Type: new Abstract: LLM-based agents for text-to-SQL often struggle with latency-performance trade-off, where performance ...
The Power of Power Law: Asymmetry Enables Compositional Reasoning
arXiv:2604.22951v1 Announce Type: new Abstract: Natural language data follows a power-law distribution, with most knowledge and skills appearing at ve...
On the Existence of an Inverse Solution for Preference-Based Reductions in Argumentation
arXiv:2604.22958v1 Announce Type: new Abstract: Preference-based argumentation frameworks (PAFs) extend Dung's approach to abstract argumentation (AAF...
Towards Causally Interpretable Wi-Fi CSI-Based Human Activity Recognition with Discrete Latent Compression and LTL Rule Extraction
arXiv:2604.22979v1 Announce Type: new Abstract: We address Human Activity Recognition (HAR) utilizing Wi-Fi Channel State Information (CSI) under the ...
FormalScience: Scalable Human-in-the-Loop Autoformalisation of Science with Agentic Code Generation in Lean
arXiv:2604.23002v1 Announce Type: new Abstract: Formalising informal mathematical reasoning into formally verifiable code is a significant challenge f...
A Systematic Approach for Large Language Models Debugging
arXiv:2604.23027v1 Announce Type: new Abstract: Large language models (LLMs) have become central to modern AI workflows, powering applications from op...
A Decoupled Human-in-the-Loop System for Controlled Autonomy in Agentic Workflows
arXiv:2604.23049v1 Announce Type: new Abstract: AI agents are increasingly deployed to execute tasks and make decisions within agentic workflows, intr...
Don't Make the LLM Read the Graph: Make the Graph Think
arXiv:2604.23057v1 Announce Type: new Abstract: We investigate whether explicit belief graphs improve LLM performance in cooperative multi-agent reaso...
Analytica: Soft Propositional Reasoning for Robust and Scalable LLM-Driven Analysis
arXiv:2604.23072v1 Announce Type: new Abstract: Large language model (LLM) agents are increasingly tasked with complex real-world analysis (e.g., in f...
Towards Automated Ontology Generation from Unstructured Text: A Multi-Agent LLM Approach
arXiv:2604.23090v1 Announce Type: new Abstract: Automatically generating formal ontologies from unstructured natural language remains a central challe...
PhySE: A Psychological Framework for Real-Time AR-LLM Social Engineering Attacks
arXiv:2604.23148v1 Announce Type: new Abstract: The emerging threat of AR-LLM-based Social Engineering (AR-LLM-SE) attacks (e.g. SEAR) poses a signifi...
Judging the Judges: A Systematic Evaluation of Bias Mitigation Strategies in LLM-as-a-Judge Pipelines
arXiv:2604.23178v1 Announce Type: new Abstract: LLM-as-a-Judge has become the dominant paradigm for evaluating language model outputs, yet LLM judges ...
From Coarse to Fine: Self-Adaptive Hierarchical Planning for LLM Agents
arXiv:2604.23194v1 Announce Type: new Abstract: Large language model-based agents have recently emerged as powerful approaches for solving dynamic and...
StoryTR: Narrative-Centric Video Temporal Retrieval with Theory of Mind Reasoning
arXiv:2604.23198v1 Announce Type: new Abstract: Current video moment retrieval excels at action-centric tasks but struggles with narrative content. Mo...
Discovering Agentic Safety Specifications from 1-Bit Danger Signals
arXiv:2604.23210v1 Announce Type: new Abstract: Can large language model agents discover hidden safety objectives through experience alone? We introdu...
AdaMamba: Adaptive Frequency-Gated Mamba for Long-Term Time Series Forecasting
arXiv:2604.23239v1 Announce Type: new Abstract: Accurate long-term time series forecasting (LTSF) requires the capture of complex long-range dependenc...
CAP-CoT: Cycle Adversarial Prompt for Improving Chain of Thoughts in LLM Reasoning
arXiv:2604.23270v1 Announce Type: new Abstract: Chain-of-Thought (CoT) prompting has emerged as a simple and effective way to elicit step-by-step solu...
Active Inference: A method for Phenotyping Agency in AI systems?
arXiv:2604.23278v1 Announce Type: new Abstract: The proliferation of agentic artificial intelligence has outpaced the conceptual tools needed to chara...
LEGO: An LLM Skill-Based Front-End Design Generation Platform
arXiv:2604.23355v1 Announce Type: new Abstract: Existing LLM-based EDA agents are often isolated task-specific systems. This leads to repeated enginee...
GSAR: Typed Grounding for Hallucination Detection and Recovery in Multi-Agent LLMs
arXiv:2604.23366v1 Announce Type: new Abstract: Autonomous multi-agent LLM systems are increasingly deployed to investigate operational incidents and ...
Constraint-Based Analysis of Reasoning Shortcuts in Neurosymbolic Learning
arXiv:2604.23377v1 Announce Type: new Abstract: Neurosymbolic systems can satisfy logical constraints during learning without achieving the intended c...
SoccerRef-Agents: Multi-Agent System for Automated Soccer Refereeing
arXiv:2604.23392v1 Announce Type: new Abstract: Refereeing is vital in sports, where fair, accurate, and explainable decisions are fundamental. While ...
When Corrective Hints Hurt: Prompt Design in Reasoner-Guided Repair of LLM Overcaution on Entailed Negations under OWL~2~DL
arXiv:2604.23398v1 Announce Type: new Abstract: We report a reproducible error pattern in GPT-5.4 on OWL~2~DL compliance queries: the model frequently...
IndustryAssetEQA: A Neurosymbolic Operational Intelligence System for Embodied Question Answering in Industrial Asset Maintenance
arXiv:2604.23446v1 Announce Type: new Abstract: Industrial maintenance environments increasingly rely on AI systems to assist operators in understandi...
ArguAgent: AI-Supported Real-Time Grouping for Productive Argumentation in STEM Classrooms
arXiv:2604.23449v1 Announce Type: new Abstract: Argumentation is a core practice in STEM education, but its productivity depends on who participates a...
Ulterior Motives: Detecting Misaligned Reasoning in Continuous Thought Models
arXiv:2604.23460v1 Announce Type: new Abstract: Chain-of-Thought (CoT) reasoning has emerged as a key technique for eliciting complex reasoning in Lar...
Escher-Loop: Mutual Evolution by Closed-Loop Self-Referential Optimization
arXiv:2604.23472v1 Announce Type: new Abstract: While recent autonomous agents demonstrate impressive capabilities, they predominantly rely on manuall...
Agentic Adversarial Rewriting Exposes Architectural Vulnerabilities in Black-Box NLP Pipelines
arXiv:2604.23483v1 Announce Type: new Abstract: Multi-component natural language processing (NLP) pipelines are increasingly deployed for high-stakes ...
Do Transaction-Level and Actor-Level AML Queues Agree? An Empirical Evaluation of Granularity Effects on the Elliptic++ Graph
arXiv:2604.23494v1 Announce Type: new Abstract: Graph-based anti-money laundering (AML) systems on blockchain networks can score suspicious activity a...
MetaGAI: A Large-Scale and High-Quality Benchmark for Generative AI Model and Data Card Generation
arXiv:2604.23539v1 Announce Type: new Abstract: The rapid proliferation of Generative AI necessitates rigorous documentation standards for transparenc...
FinGround: Detecting and Grounding Financial Hallucinations via Atomic Claim Verification
arXiv:2604.23588v1 Announce Type: new Abstract: Financial AI systems must produce answers grounded in specific regulatory filings, yet current LLMs fa...
When AI reviews science: Can we trust the referee?
arXiv:2604.23593v1 Announce Type: new Abstract: The volume of scientific submissions continues to climb, outpacing the capacity of qualified human ref...
Thinking Like a Clinician: A Cognitive AI Agent for Clinical Diagnosis via Panoramic Profiling and Adversarial Debate
arXiv:2604.23605v1 Announce Type: new Abstract: The application of large language models (LLMs) in clinical decision support faces significant challen...