Scientific Research
Search documents
OpenAI发布权威AI科研基准,扯下AI遮羞布:奥赛金牌≠一流科学家
3 6 Ke· 2025-12-17 09:00
Core Insights - OpenAI has released a new benchmark called FrontierScience to evaluate AI's scientific reasoning capabilities in physics, chemistry, and biology, revealing that AI still has a long way to go to match true scientists [1][6][17] Group 1: Benchmark Design and Structure - FrontierScience consists of over 700 text-based questions, including 160 "Gold Set" questions, with 100 competition-style questions and 60 original research sub-tasks designed by PhD-level researchers [9][12] - The competition track emphasizes short-answer formats for easy verification, while the research track uses a 10-point scoring system, requiring at least 7 points to pass [9][12] - The quality of questions is ensured through collaboration with 42 international award winners and 45 qualified scientists across various fields [11][12] Group 2: AI Performance and Comparison - Initial testing showed that GPT-5.2 scored 77% on competition questions and 25% on research questions, leading the pack, while Gemini 3 Pro followed closely with 76% on competition questions [13] - In a previous benchmark, GPT-4 scored only 39% on a question set designed by PhD experts, significantly lower than the expert baseline of 74% [6][12] Group 3: Challenges and Limitations - OpenAI acknowledges that advanced models still make reasoning, logic, and factual errors, and that longer processing times often correlate with higher accuracy [15][17] - FrontierScience is designed to standardize assessments but does not evaluate the models' ability to generate truly novel hypotheses or interact with multimodal data and real-world experimental systems [17] Group 4: Future Directions - OpenAI plans to iterate on the question bank, expand the fields covered, and include more real-world assessments to determine the practical impact of these systems on scientific work [17]
北大团队最新研究:AI零基础自主推导出牛顿第二定律
Feng Huang Wang· 2025-12-10 04:31
凤凰网科技讯 12月10日,人工智能在科研中扮演的角色正在发生质变。北京大学物理学院团队近日展 示了一个具备自主理论构建能力的AI系统,并因成功从实验数据中独立推导出经典力学定律而登上 《自然》杂志。该成果被视为AI从辅助工具迈向"可独立认知的科研主体"的重要信号。 据介绍,北大团队研发的"AI-牛顿"系统在没有预设任何物理背景知识的条件下,被置于包含小球、弹 簧等装置的46组物理实验数据中。它不仅能识别实验中的运动模式,更在迭代推理过程中逐步构建 出"力""质量""加速度"等基础概念,并最终得出牛顿第二定律F=ma。整个过程依靠的是一种近似于科 学家"先假设、再求证"的合情推理策略,而非传统大模型的简单拟合。 研究团队将实验装置与数据生成器预置于"实验库",再让系统在"理论库"中不断积累自身构建的概念与 规律。三层结构(符号、概念、定律)让 AI 得以在繁杂数据与可解释公式之间建立联结,使推导路径 可回溯、逻辑链条清晰。这也是该系统突破神经网络"黑箱"困境的关键。 外界普遍关注这项尝试对未来科研方式的潜在影响。《自然》引用哈佛大学学者的观点指出,这种编程 方式能够鼓励模型主动逼近核心概念,或将成为未来科学发 ...
X @Cointelegraph
Cointelegraph· 2025-11-16 04:00
🔥 INSIGHT: Blockchain treasuries are creating fast, decentralized funding pipelines for scientific research. https://t.co/2K8UVbnyE9 ...
Harvard scientist: U.S. risks falling behind China in innovation #shorts
60 Minutes· 2025-11-11 18:01
So this is your lab. >> Yes, this is where we do the organ chip cultures. >> Ingber relies largely on federal funding for his work identifying new therapies using what he calls organs on chips.This breakthrough technology, tiny tissue line devices can replace animal testing. So what does that allow you to do. We can study how the body normally works.We study response to drugs, response to toxins, drug delivery systems. >> Federal grants have made up almost half the university's research funding. Ingber and ...
华人学者一天发表了11篇Nature论文
生物世界· 2025-10-16 08:25
Core Insights - On October 15, 2025, a total of 22 papers were published in the prestigious journal Nature, with several authored by Chinese scholars, highlighting the significant contributions of Chinese researchers in various scientific fields [2][4][6][9][12][21][22][24]. Group 1: Research Contributions - Zhang Zhenhua from RWTH Aachen University published a paper on "Deaminative cross-coupling of amines by boryl radical β-scission," focusing on a novel reaction mechanism involving boryl radicals [2]. - A collaborative work by Xu Minyu from National University of Singapore and Zhang Xinglong from Chinese University of Hong Kong discussed "Photocatalytic oxygen-atom transmutation of oxetanes," presenting advancements in photocatalysis [4]. - A team from Shanghai Jiao Tong University, Fudan University, and Nanjing University of Science and Technology introduced "Tin-based perovskite solar cells with a homogeneous buried interface," contributing to solar energy technology [6]. - The Taiwan Precision Medicine Initiative was highlighted in a paper by researchers from Academia Sinica, providing a cohort for large-scale studies in precision medicine [9]. - A study on "Population-specific polygenic risk scores for people of Han Chinese ancestry" was published by researchers from Academia Sinica, focusing on genetic risk assessment [12]. Group 2: Environmental and Material Science - Lin Yucheng from Rutgers University published a paper on "Modern sea-level rise breaks 4,000-year stability in southeastern China," addressing significant environmental changes [17]. - Research on "Patchy nanoparticles by atomic stencilling" was conducted by Chen Qian from the University of Illinois, exploring innovative methods in nanotechnology [19]. - A study on "A conserved H3K14ub-driven H3K9me3 for chromatin compartmentalization" was published by researchers from East China Normal University and the Chinese Academy of Sciences, contributing to the understanding of chromatin biology [21]. - Tsinghua University’s Fang Lu published a paper on "Integrated lithium niobate photonics for sub-ångström snapshot spectroscopy," advancing photonic technologies [22]. - Chen Chi-Fang from UC Berkeley introduced "Efficient quantum thermal simulation," focusing on quantum computing applications [24].
X @The Economist
The Economist· 2025-10-11 18:40
Scientific Research Focus - This year's Nobel-winning scientific research has potential uses ranging from cancer treatments to extracting water from dry desert air [1]
X @The Economist
The Economist· 2025-10-11 13:00
Scientific Research Focus - This year's Nobel-winning scientific research has potential uses ranging from cancer treatments to extracting water from dry desert air [1] Industry Application - The research could be applied to cancer treatments [1] - The research could be applied to extracting water from dry desert air [1]
‘I haven’t been able to do my job since January’: Furloughed NIH worker speaks out
MSNBC· 2025-10-03 21:16
ing and joining us now is one of the hundreds of thousands of furled government workers. Jenna Norton is a program director at the NIH, the National Institutes of Health. She's now joining us in her personal capacity.Uh so this is not your first shutdown, but you spoke with my colleague Shannon Petty Piece, and you said you had a lump in your throat as you left the NIH offices this week. Why why this time did it feel different. You know, I think this shutdown is just completely different from anything that ...
'Burning Our Future': MIT’s Rafael Reif Warns About Government Research Cuts
Bloomberg Television· 2025-09-21 12:00
Westin: This is a story about trees not growing to the sky, even if those trees are holding up an important part of the U.S. economy. President Trump has made no secret of his quest to cut funding to some of the nation's most prestigious universities. - We want money to go to all universities, not Harvard.There have $52 billion. - Harvard's voice is loud and clear. Westin: To date, much of the discussion has centered on the politics of it all, whether higher education has gotten too "woke" and whether it's ...
Former MIT president warns on research funding cuts #shorts #mit #highereducation #funding #cuts
Bloomberg Television· 2025-09-20 17:07
Research & Funding Concerns - Scientific research funding cuts are detrimental to the future, potentially halting crucial research [1] - Paused research leads to laboratory closures, loss of experienced personnel, and a decline in expertise [2] - Federal funding is a critical component of the research dimension within the innovation economy [4] Startup Ecosystem & Market Conditions - Despite challenges, the Boston startup ecosystem showed strength in the first half of 2025 [2] - Healthcare and biotech research has been particularly affected by pauses and job cuts, creating volatility [3] - Deal flow has remained consistent, with exceptional teams and technological advancements still emerging [3][4]