资讯列表
Ant Technology Group Accelerates AI Strategy and Establishes AI+ Industry Innovation Lab
On June 10, at the 9th Digital Finance and Financial Security Conference in 2025, Ant Tech announced the comprehensive upgrade of the Taurus Lab. Originally named the "Digital Identity Security Lab," it has been upgraded to the "Artificial Intelligence Industrial Innovation" Lab. The lab will focus on key technological breakthroughs in the application of AI large models in industrial scenarios.
6/16/2025 9:48:51 AM
AI在线
Deep Integration of Grok AI into X Algorithm: Small Creators Welcome Spring with Smarter Comment Interaction
X platform recently announced a major update to its recommendation algorithm, deeply integrating its self-developed Grok AI technology, with the aim of optimizing user content experience and significantly increasing the exposure opportunities for small accounts. This adjustment is considered by the industry as an important innovation in the traditional social media content distribution mechanism.Core Upgrade of the Algorithm: Quality First, Scale Makes WayThe core of the latest algorithm update lies in providing more visibility to small accounts. The platform now highlights content from lesser-known creators, creating a more equitable competitive environment.
6/16/2025 9:48:51 AM
AI在线
Former DeepSeek executive secretly starts new AI Agent project, already backed by top VC
According to the news from Tiger嗅, a core executive of the domestic large model company DeepSeek has quietly left and started a new business half a year ago, and plans to launch its first Agent product around Christmas in 2025.Sources close to the matter told Tiger嗅 that the executive once held the role of "CTO" at DeepSeek. However, some insiders pointed out that DeepSeek's internal structure does not clearly set up a "CTO" position. In name, this position may not exist, but there is indeed an executive who takes on technical coordination and R&D decision-making responsibilities similar to those of a CTO.Reliable sources also revealed that this startup project has already received financing support from a leading VC, with the specific amount undisclosed.
6/16/2025 9:48:51 AM
AI在线
How is LVMH using AI technology to cope with the downturn in the luxury goods market?
As the luxury market faces challenges of slowing growth, LVMH is turning its attention to artificial intelligence (AI) technology in an effort to enhance operational efficiency and customer retention. The group, which owns renowned luxury brands such as Tiffany, Dior, and Celine, recently revealed that AI will play a crucial role in their response to market changes.Over the past four years, LVMH has collaborated with Google Cloud to build a centralized data platform that consolidates data from its 75 brands. They are now applying predictive AI, generative AI, and intelligent agents across various areas, including supply chain planning, pricing, product design, marketing, and personalized experiences.
6/16/2025 9:48:51 AM
AI在线
Grok Takes Over Twitter Algorithm! Small Accounts Can Also Go Viral, Is This the Golden Era for Newcomers to Build Their Accounts?
According to the latest news, Grok, the artificial intelligence developed by xAI, has officially taken over the recommendation algorithm of X. This change will not only reshape users' content experience but also bring unprecedented opportunities to small accounts with fewer fans and newcomers. Below is the latest information compiled by AIbase for you to gain a deeper understanding of the details of this transformation and its potential impact.Grok-Driven Recommendation Algorithm: Small Account Content Gets Priority RecommendationsGrok, as the core AI product of xAI, has been deeply integrated into X's recommendation system this time with the aim of optimizing the accuracy and fairness of content distribution.
6/16/2025 9:48:51 AM
AI在线
All Ohio State University students will receive AI training to enhance technical application skills
Ohio State University recently announced that, starting this year, all students will be required to receive training related to artificial intelligence (AI). The goal is to enable students to effectively combine traditional learning with AI technology. University president Walter Carter Jr.
6/16/2025 9:48:51 AM
AI在线
NVIDIA and HKU Collaborate to Launch New Visual Attention Mechanism, Boosting High-Resolution Generation Speed by Over 84 Times!
Recently, The University of Hong Kong and NVIDIA jointly developed a new visual attention mechanism called Generalized Spatial Propagation Network (GSPN), which has made significant breakthroughs in high-resolution image generation.Although traditional self-attention mechanisms have achieved good results in natural language processing and computer vision fields, they face dual challenges of huge computational overhead and loss of spatial structure when handling high-resolution images. The computational complexity of traditional self-attention mechanisms is O(N²), making it very time-consuming to process long contexts.
6/16/2025 9:48:51 AM
AI在线
Thinkless框架:让LLM学会“聪明偷懒”的智慧
大家好,我是肆〇柒。 今天,我想和大家聊一下,我看到关于自适应思考的另外一片论文,它介绍了Thinkless 框架,并且还有开源仓库。 今天我们要了解的 Thinkless 这个框架,由新加坡国立大学的研究人员提出,它能够巧妙地解决当前推理语言模型(LLM)在处理简单问题时过度推理、浪费资源的难题。
6/16/2025 9:40:48 AM
肆零柒
React 正式接入 AI,你知道吗?
最近,React 官方在其 GitHub 仓库中引入了一个实验性项目 —— React MCP Server,通过 MCP 将大模型与 React 编译工具链连接起来。 简单来说,它使 AI 助手不仅能够生成代码,还能以“工具调用”的方式直接参与 React 代码的编译优化和文档查询。 MCP Server 实现了 MCP 协议,并定义了两个主要工具供 AI 使用:编译工具:通过集成 babel-plugin-react-compiler,该工具允许 AI 对传入的 React 组件代码进行编译优化。
6/16/2025 9:36:18 AM
CUGGZ
迈向人工智能的认识论:涌现能力和思路链的忠实性
一、涌现能力:真实现象还是测量幻象过去几年中一个有趣的观察是,随着我们扩大模型规模(在参数和训练数据方面),它们开始展现出较小模型所不具备的全新定性能力。 Wei 等人(2022 年)将这些能力称为涌现能力,并将这种能力定义为“如果某种能力在较小模型中不存在,但在较大模型中存在,则为涌现能力”,以至于“不能简单地通过推断较小模型的性能来预测它”。 换句话说,这种技能不是平稳的改进,而是在模型超过某个规模阈值后突然出现。
6/16/2025 9:28:09 AM
晓晓
全球首次,Transformer「混血」速度狂飙65倍!英伟达已下注
扩散方法打入语言模型领域! 最近,康奈尔博士生Subham Sahoo,在X介绍了扩散大语言模型的最新工作。 这项研究引发了AI研究领域的思考。
6/16/2025 9:13:00 AM
12年博士研究,AI两天爆肝完成!科研效率狂飙3000倍,惊动学术圈
多伦多大学、哈佛MIT等机构联手AI,短短2天内,竟干完了科学家12年的活! 研究一出,在全网掀起了巨震。 沃顿商学院CS教授Ethan Mollick大受震撼,「AI综述论文体量越来越大,而准确性超越了人类」。
6/16/2025 9:12:00 AM
破解三大数据库AI难题!北大&亚马逊推出全球首个图中心RDB基础模型
在企业系统和科学研究中普遍存在、结构复杂的关系型数据库(Relational DataBase, RDB)场景中,基础模型的探索仍处于早期阶段。 这是因为RDB中的多表交互和异质特征,使传统通用大模型在此类结构化环境下难以直接发挥效能。 基于此,北京大学张牧涵团队联合亚马逊云科技共同提出了Griffin:一个具有开创性的、以图为中心的RDB基础模型。
6/16/2025 9:10:00 AM
刚刚!陶哲轩3小时对话流出:AI抢攻菲尔兹奖倒计时
昨天,陶哲轩公开表示:在未来,AI有可能获得菲尔兹奖。 在时长3小时14分钟的节目中,他还讨论了数学和物理学中最难的问题:纳维–斯托克斯问题 万有理论 广义相对论 孪生素数猜想 费马大定理 考拉兹猜想 P=NP 问题一开始,陶哲轩就认为统一物理理论(比如量子力学与广义相对论)之所以难以进展,是因为还没有找到正确的数学语言。 而AI可能帮助人类实现「大统一理论」,至少也是发现过程中的一部分。
6/16/2025 9:07:00 AM
全球程序员炸锅!老黄伦敦放豪言:编程语言的未来是「Human」
你知道人类目前的编程语言到底有多少种? C、C 、C#、Java、Python、PHP、JavaScript、Go、Rust……但不论有多少种,这些都不重要了。 未来也许只会剩下唯一一种编程语言,AI将彻底改写人机交互的定义。
6/16/2025 9:03:00 AM
AI改变世界!微软发布700个真实AI Agent、Copilot应用案例
我们经常会听到AI改变世界的观点,但除了问答、查询、总结之外,似乎很少会看到其他真实的AI应用案例。 所以,微软作为全球AI领导者之一,发布了700个真实的Agent智能体、Microsoft Copilot应用案例,来帮助大家真实地了解到底AI是如何改变我们的工作模式。 这些客户有世界500强企业,有各领域的独角兽,也有世界常青藤名校,业务范围涵盖金融、医疗、科技、教育、汽车制造、零售、电商、能源、航空航天等众多领域。
6/16/2025 9:02:00 AM
比马斯克Neuralink工艺更先进!我国成功开展首例侵入式脑机接口临床试验
用意念操控万物的“脑机接口”,就这样离我们更近了? 据央妈最新报道,我国已成功开展首例侵入式脑机接口临床试验——一位因意外事故导致四肢截肢的男子,现在仅凭意念就能玩五子棋游戏、发短信等等。 这项研究由中国科学院脑科学与智能技术卓越创新中心,联合复旦大学附属华山医院以及相关企业合作完成。
6/16/2025 9:00:00 AM
AI ASMR突然火爆全网!3天狂揽近10万粉丝,一条切水果视频播放量破1650万
短视频平台熬夜冲浪的我突然发现,AI ASMR突然火了? ! 大家可以打开声音欣赏一下这条切割浆果的视频,它拿下了16500000的浏览量。
6/16/2025 8:54:00 AM