AIoT研习社 · 2026-05-01👍 👎

大模型竞赛的焦点,正在从'参数规模'转向'推理能力'.这不... 推理性能:智能的真正试金石发布会上的演示令人印象深刻.一个... [DeepSeek V4's launch signals a shift in the large model race from parameter scale to reasoning performance, with impressive demos highlighting inference as the new benchmark for AI intelligence.]

梁子智算圈 · 2026-05-01👍 👎

发布中国首个对标业界最先进主流模型的旗舰大模型.＂ 这句话的重量,很多弱电人可能还没感受到. 梁子跟你说: 这是国内AI生... [iFlytek announces plans to release China's first flagship large model benchmarked against leading global models on Huawei's Ascend 950 platform in October.]

施显Works · 2026-05-01👍 👎

简言之,这个工作发现,当前最强的大模型(包括GPT-5.5)在真实生活文本的平均正确率仅约 22%,说明现有的AI在处理日常上下文... [Tencent's CL-Bench Life benchmark reveals that even the strongest current LLMs, including GPT-5.5, achieve only about 22% accuracy on real-life text tasks, highlighting major gaps in everyday contextual understanding.]

Python数值计算 · 2026-05-01👍 👎

一、概述2026年4月16日至24日,全球至少8款大模型密集发布或开源.本月主要动态包括:开源与闭源差距收窄:开源模型在部分编... [A roundup of April 2026 LLM developments notes at least 8 models released or open-sourced globally between April 16–24, with narrowing gaps between open- and closed-source models and growing Agent commercialization.]

亮虾哥 · 2026-05-01👍 👎

2026年4月大模型发布时间轴可视化图 ❷为什么R1惊艳,V4平平?腾讯新闻的分析里有个关键对比:对比维度R1时期(2025年初)V... [An analysis of April 2026's crowded LLM release calendar examines why DeepSeek V4 failed to replicate the impact of R1, citing timing competition and shifting industry expectations.]

电子信息科创E站 · 2026-05-01👍 👎

04月22项目来源:AI+科创E站发布时间:2026-04-22政策解读丨国务院重磅发布!深入实施'人工智能+'行动,支持采购大模型、智... [The State Council has issued a major directive to deepen the 'AI+' action plan, including government procurement support for large language models and AI agent services, with a target of growing the AI service sector to 10 trillion yuan by 2030.]

西风冷楼阙 · 2026-05-01👍 👎

领域在政策引导、企业实践和技术创新的共同推动下,呈现出以人工智能(AI)为驱动引擎、安全底座持续升级的鲜明发展态势.AI不... [China's state-endorsed tech innovation (Xinchuang) sector is entering a concrete engineering deployment phase, with policy guidance, enterprise practice, and capital investment converging around AI as the core driver and upgraded security infrastructure.]

搜药 · 2026-05-01👍 👎

问AI就行,不用翻半天文件.说到底,文件释放的信号相当明确,药品监管正在从'人盯人'转向'数智驱动'.02从医保基金到执业... [China's National Drug Administration has issued a directive signaling a comprehensive shift in pharmaceutical regulation from human-monitored oversight to AI and data-intelligence-driven systems, affecting pharmacies, medical insurance funds, and licensed professionals.]

成都日报 · 2026-05-01👍 👎

据网信中国4月30日消息,中央网信办部署开展为期4个月的'清朗·整治AI应用乱象'专项行动,将重点整治不当使用AI'复活逝者'... [China's Cyberspace Administration (CAC) has launched a 4-month 'Qinglang' campaign targeting 14 categories of AI application abuses, with a key focus on cracking down on the commercialized and unauthorized AI 'resurrection' of deceased individuals.]

律师蒋法 · 2026-05-01👍 👎

4月30日,中央网信办印发通知,在全国范围内部署开展为期4个月的 '清朗·整治AI应用乱象' 专项行动.此次行动旨在规范AI服务... [On April 30, the CAC issued a nationwide notice launching a 4-month 'Qinglang' special campaign to regulate AI services, targeting 14 specific categories of misconduct including deepfake face theft, data poisoning, and unauthorized account management.]

乐享直供 · 2026-05-01👍 👎

下一步:人工智能+软件专项行动工信部明确:培育模型即服务、智能体即服务新业态加快智能编程研发应用推动基础软件、工业软件... [China's Ministry of Industry and Information Technology has launched an AI+Software special action plan to cultivate model-as-a-service and agent-as-a-service ecosystems, accelerate intelligent programming, and drive AI integration into foundational and industrial software.]

转型智道 · 2026-05-01👍 👎

4月28日,工信部、国家数据局联合印发通知,正式启动2026年'模数共振'行动,这是'人工智能+制造'专项行动的核心攻坚举措... [On April 28, China's MIIT and National Data Administration jointly issued a notice officially launching the 2026 'Model-Data Resonance' initiative, a flagship effort under the 'AI+Manufacturing' special action plan targeting 20 major industrial sectors.]

十师北屯市生态环境局 · 2026-05-01👍 👎

国家数据局综合司2026年4月3日附件数据产权登记工作指引(试行)(公开征求意见稿)第一章 总则第一条 制定目的为建立健全数据... [China's National Data Administration is seeking public feedback on draft guidelines for a data property rights registration system, aimed at establishing a comprehensive framework for data ownership and commercialization.]

人工智能伦理 · 2026-05-01👍 👎

(征求意见稿),指出人工智能应用伦理安全风险,提出伦理安全原则和伦理安全指引.1、伦理安全风险· 人类主导权弱化:可能导... [A draft AI application ethics and safety guideline document is set for release, identifying key ethical risks including the weakening of human control and proposing corresponding safety principles.]

风驰 · 2026-05-01👍 👎

多位人大代表和政协委员提交了关于加快制定《人工智能法》的议... 完善AI伦理准则,禁止算法歧视、算法压榨等损害劳动者权益的行... [Multiple NPC and CPPCC delegates have submitted proposals to accelerate legislation on an AI Law, calling for ethics guidelines that ban algorithmic discrimination and exploitation of workers.]

AI资讯 · 2026-05-01👍 👎

近日,工信部等十部门联合印发《人工智能科技伦理审查与服务办法(试行)》,这是中国首部专门针对AI活动的全流程伦理审查制度... [Ten ministries led by MIIT have jointly issued China's first full-process AI ethics review regulation, covering all AI activities under a binding trial framework, with a comparative look at EU and US medical AI governance.]

高校思政研习社 · 2026-05-01👍 👎

'人工智能+教育' 国家行动来了!《'人工智能+教育'行动计划》解读前言当前,人工智能对教育底层逻辑和样态重塑带来系统性... [China's 'AI + Education' Action Plan has been released, signaling a national-level initiative to systematically reshape education through artificial intelligence.]