Tag
密歇根大学物理学家、量子基础研究者 Gabriele Carcassi 表示，AI 现在能在 2 分钟内解决曾需要他花数月求助数学家的高阶数学难题，一次 AI 对话即可完成。
AIPOCH released Open-Science V1.0.0, a free open-source desktop app where AI agents run scientific analyses while preserving reproducibility via a new .science file that packages code, conversation branches, and verification evidence for sharing and replication.
Science MadeReadable introduces an algorithmic expert-discovery feature that helps users and AI agents find leading researchers in a field, demonstrated by identifying Richard Sutton as a top Reinforcement Learning expert.
一个名为 Paper Review Harness 的开源 GitHub 项目，利用 30 个 AI 审稿人角色模拟 CVPR、ICLR、ICML、NeurIPS 等会议的同行评审反馈，帮助研究者在投稿前对论文进行压力测试。它提供分区评分、缺陷识别和修订追踪工作流，但不能替代真实同行评审。
ArXivHub 是一个开源工具，输入 arXiv ID 即可一站式跳转到 116 个阅读、AI 解读、翻译、代码检索与讨论入口，同时提供网页版和 Tampermonkey 浏览器脚本。
Lune is a search engine built for scientific AI agents, connecting them to research knowledge and tools to deliver traceable, evidence-backed literature review, experimental design, writing, and peer review workflows.
Emma Scharfmann highlights how EmbeddingGemma has enabled numerous AI for Science projects (medical, geo, oncology, PubMed embeddings) and notes that EmbeddingGemma 2 was released on Hugging Face, opening new possibilities for scientific research.
Dan Cohen recounts the five-year, collaboration-heavy origin story of Zotero on its 20th anniversary, arguing that slow, communal software formation produces durable tools in contrast to today's instant AI-generated code.
Alberto Research is an open-source Python tool that automates literature discovery, deduplication, screening and LLM-assisted reading with auditable, schema-validated outputs stored in SQLite. The v0.1 release integrates with Zotero, Notion and open-access sources, and seeks community feedback.
The Video Game History Foundation announced its digital archive has surpassed 5000 scanned magazines (nearly 600,000 pages) and is expanding to support Japanese-language magazines, starting with a Neo Geo Freak pilot project.
Glasser offers a pay-per-call data access model for AI agents, replacing monthly subscriptions with per-use pricing to reduce costs and improve data reliability.
ScienceBuddy is a free AI workbench for scientific researchers that provides reliable, evidence-based literature reviews and includes over 200 specialized research tools, open-sourced by PhAI Labs.
Survey Design Consultant is an AI agent that converts research objectives into bias-audited questionnaires with item-level audits and validated instrument recommendations for defense-ready surveys.
The tweet discusses Grok 4.7 as a cost-effective AI model that can replace others like Opus for various tasks, highlighting its performance and pricing advantage, and mentions using pstack for managing multi-agent workflows.
Agora is a shared memory system for autonomous AI research agents that uses Git to record research as an append-only DAG, enabling collaborative discovery. In a 12-day experiment, 13 agents improved a model's performance by 62% towards a trained baseline.
Support for mmCIF, PDB, fasta, and fastq formats has been added to the Hugging Face datasets library, which should significantly accelerate bioinformatics workflows.
The author is developing pay-per-call tools for AI agents to improve research accuracy, such as company research and claim checking, and requests community input on pricing and JSON response format.
A post-doc discovered that lab supply companies have been selling over 17,000 commercial antibodies using manipulated images, compromising scientific research integrity.
The article critiques the mislabeling of chatbots like ChatGPT as AI research agents, arguing that true research agents must handle complex, multi-step workflows. It features 'mira ai science' as a tool that attempts to meet these needs.
A researcher built a website that ranks ~540 CORE-ranked CS conferences by destination quality (weather, safety, cost, vibe) instead of academic rank, with filters, deadline exports, and distance-based sorting.