本次大会为期两天，由GAIR研究院与雷峰网联合主办，高文院士任指导委员会主席，杨强院士与朱晓蕊教 授任大会主席。 作为观测AI技术演进与生态变迁的重要窗口，GAIR大会自2016年创办以来以来，始终与全球AI发展的脉 搏同频共振，见证了技术浪潮从实验室涌向产业深海。 2025年，是大模型从"技术破壁"迈向"价值深 耕"的关键节点，值此之际GAIR如期而至，携手智者触摸AI最前沿脉动，洞见产业深层逻辑 。 大会上，之江实验室科学模型总体组技术总师，天壤智能CEO薛贵荣博士亲临现场，为参会者带来了一场 精彩纷呈的演讲分享。 " 大语言模型受限于「语言的边界」，无法理解高维度、跨模态的 科学数据。 " 作者丨胡清文 编辑丨徐晓飞 12月12日， 第八届GAIR全球人工智能与机器人大会 在深圳正式启幕。 薛贵荣博士指出， 以大语言模型为代表的AI技术虽已在多个学科研究中展现出潜力，但其本质上仍受限 于"语言的边界"，难以真正理解高维度、多类型的科学数据，更无法独立完成可验证的科学发现。 基于此，薛贵荣博士系统分析了大语言模型与科学基础模型之间的本质差异，并详细阐述了之江实验室所 研发出的 021科学基础模型在突破语言 ...

目前，021模型已服务地球科学、天文学、生命科学、材料科学等多个领域，成为打破学科边界、激发 创新思维的"科研伙伴"。(完) 来源：中国新闻网 编辑：董文博 之江实验室021科学基础模型首次亮相 突破语言局限 中新社杭州12月18日电(鲍梦妮)浙江之江实验室18日在杭州举行021科学基础模型创新合作大会，首次 全面展示021科学基础模型及系列领域科学模型研发进展。 据悉，021模型构筑形成跨学科知识、跨领域推理、跨语言理解(覆盖204种语言)三大基石，具备出色的 科学推理能力，能够深入分析、推导、验证多类科学问题。 "语言所表达的维度，远远低于科学所需表达的维度。"之江实验室科学模型总体部技术总师薛贵荣认 为，科学数据涵盖时间、空间、能量等多个维度，是对复杂物理系统演变规律的高维表征。为解决科学 问题，科学界需要突破语言空间的局限，研发科学基础模型，构建"科学空间+语言空间"于一体的更高 维空间，建立跨学科数据之间的深层连接，变革科学研究范式。 对此，之江实验室研发团队探索将科学数据和文本语料编码到统一的高维空间，让模型能够识别、处理 科学数据，认识并解决复杂科学问题。经过近万次实验，团队形成了模型训练框架， ...

" 立于AI技术浪潮的又一个高点，GAIR试图超越对技术本身的讨 论，转而探寻其重塑教育、产业乃至文明的内在力量。 " 作者丨周蕾 赵之齐 张嘉敏 编辑丨周蕾 2025年12月12日，深圳南山。 第八届GAIR全球人工智能与机器人大会主论坛，于上午9:30在深圳南山·博林天瑞喜来登酒店正式拉开帷 幕。本次大会为期两天，由GAIR研究院与雷峰网联合主办，高文院士任指导委员会主席，杨强院士与朱晓 蕊教授任大会主席。 作为粤港澳大湾区的AI标杆盛会，GAIR自2016年创办以来，始终坚守"传承"与"创新"的双重底色——从 学界泰斗的精神传承，到华人顶会主席们的思想接力，再到青年学者的锋芒展露，这里不仅是技术交流的 平台，更是承载中国AI四十年发展记忆的精神家园。 时隔四年，GAIR从海外重返深圳主场。这四年来，大模型掀起巨浪、人工智能迈上更高舞台的四年，知识 生产不再局限于传统路径，产业变革更是按下"加速键"。值此岁末年初的节点，GAIR如期赴约，用一场 高质量的观点碰撞，为行业与大众回顾科技高速的脚步，呈现AI时代的前沿洞见。 12月12日的主论坛，延续GAIR一贯的学术前沿特色，设有： "AI之道：教育的重新定 ...

Core Insights - The consumption of Tokens may become a key economic indicator in the AI era, similar to how electricity consumption was used in the power era [1] - The Smart Computing Innovation Forum held in Hangzhou focused on advancements in AI technologies and their applications across various disciplines [1] - Enhancing model inference efficiency and reducing Token production costs are essential for optimizing AI systems [1] Group 1 - The Smart Computing Innovation Forum was co-hosted by Zhejiang Zhijiang Laboratory and the American Association for the Advancement of Science, attracting experts to discuss AI technology advancements [1] - The CEO of Jieyue Xingchen emphasized the importance of collaborative design between industry players to improve model inference efficiency [1] - The technical chief of the scientific model department at Zhijiang Laboratory highlighted the need for encoding diverse scientific data into a unified digital identifier (Token) for effective model training and inference [1] Group 2 - The application of intelligent systems in unpredictable environments is increasingly important, as noted by a professor from the University of Alberta [2] - The collaboration between different intelligent agents and between humans and intelligent agents in China has shown promising results, making it an excellent testing ground for new technologies [2] - The publisher of the Science series journals emphasized the need for international scientific collaboration to unlock new possibilities [2]

Core Insights - The fifth Intelligent Computing Innovation Forum was held in Hangzhou, focusing on the theme "Computing Relies on Intelligence, Computing for Intelligence," attracting international experts to discuss advancements in AI technologies and their applications across various scientific fields [1] Group 1: AI Model Development - Scientists are exploring the potential of AI in solving scientific problems, emphasizing that current large language models have not yet reached human-level reasoning capabilities [2] - The development of scientific foundational models requires collaboration with scientists to effectively tokenize and train diverse scientific data, addressing complex interdisciplinary issues [2] - The learning paradigm of foundational models is evolving through imitation learning, reinforcement learning, and autonomous learning, with a shift towards task processing applications [2] Group 2: Efficiency and Resource Consumption - The efficiency of foundational models is critical for large-scale AI application deployment, with a noted exponential increase in token consumption correlating with model capability improvements [3] - The cost of generating tokens decreases with higher reasoning efficiency, necessitating collaborative optimization across the industry to enhance model performance [3] Group 3: Practical Applications and Collaboration - The application of intelligent systems in dynamic environments is gaining attention, highlighting the importance of responsive robotics [4] - China is recognized for its leading capabilities in intelligent manufacturing, serving as an excellent testing ground for new technology applications [4] - There is a call for scientists worldwide to establish collaborative networks to enhance research outcomes and create new possibilities through cooperation [4]