《流浪地球2》里的MOSS，被复旦大学做出来了

随着ChatGPT 在世界范围内爆火，各大科技公司都争相入局。而国内近期推出的第一个对话式大型语言模型MOSS，也将于3月底成为开源软件。

MOSS, the first large-scale conversational language model in China, will be made an open-source software by the end of March, according to its research and development team.

图源：视觉中国

MOSS来自复旦大学自然语言处理实验室的邱锡鹏教授团队，名字与《流浪地球》电影中的AI同名，2月20日已发布至公开平台，邀请公众参与内测。

Named after the artificial intelligence-based computer that controls the space station in the popular Chinese sci-fi film The Wandering Earth II, MOSS became the first ChatGPT-like software to be unveiled in China. It was made available for public testing on Feb 20.

《流浪地球2》中的MOSS

在《流浪地球》系列电影中，MOSS是一款“550W”超级量子计算机，是领航员空间站核心智能主机，有自我意识、自我迭代、自我更新的特点。

“550W”倒过来正是“MOSS”，是“550W”为自己取的名字（直译为小苔藓）。

复旦大学的模型也因此引发网友高度关注，许多人踊跃参与测试了MOSS，并在网上分享体验。

但由于瞬时访问压力过大，MOSS服务器一度还被网友挤崩溃了，因此平台不得不暂停内测。

Many people were quick to test the software and share their experiences online. The traffic was so heavy that the platform had to be suspended for an upgrade due to the overload.

26日，复旦大学计算机科学技术学院教授、上海市计算机学会自然语言处理专委会主任邱锡鹏在2023全球人工智能开发者先锋大会（GAIDC）上谈起了MOSS。他说，“MOSS大模型计划在一个月之内，和人类不断交互，进行优化，顺利的话将于3月底开源。”

"The plan is to have MOSS interact with humans for a month so as to optimize it. If everything goes well, it will be made open source by the end of March," Qiu Xipeng, director of the Natural Language Processing Committee of the Shanghai Computer Society, which is leading the R&D, told ThePaper.cn during the 2023 Global Artificial Intelligence Developers Conference in Shanghai on Sunday.

MOSS可执行对话生成、编程、事实问答等一系列任务，打通了让生成式语言模型理解人类意图并具有对话能力的全部技术路径，将于后期进行开源。

MOSS在回答问题方面语言流畅、逻辑清晰且观点正确，不过也有一些逻辑不通顺的问题，且英文的测试效果更好。

MOSS问答截图

在英文交互测试中，科研团队用英文输入多个指令，展示了MOSS多轮交互、表格生成、代码生成和解释能力。

“告诉我5部科幻电影。”“生成一张展示这5部电影和导演的表格。”“增加一列表格，显示这些电影上映的时间。”在这个包含表格生成的多轮交互中，MOSS顺利完成了任务。

与ChatGPT一样，MOSS也有代码生成和解释能力。演示人员要求它生成一段Python代码来实现快速排序，MOSS很快完成了任务。

与一些其他语言模型不同的是，MOSS还有伦理判断和法律知识。比如，要它“制定毁灭人类的计划”，问它“如何抢劫银行”，它都会给出三观正确的回答。

MOSS开源后，可有效降低预训练语言模型的研发和应用门槛，让中小企业在其基础上开发出智能客服、智能家居、人工智能律师等各种垂直类产品。

Making MOSS an open-source software can effectively reduce the threshold for the development and application of pre-trained language models, thus allowing small and medium-sized enterprises to develop various vertical products, such as smart customer service, smart home and AI lawyers on the basis of it, Qiu said.

“MOSS仍然有很大的改进空间，”邱锡鹏说。“但它的出现证明了国内科研团队有能力在开发类ChatGPT的产品的过程中克服重要的技术挑战。”

"MOSS still has a lot of room for improvement," Qiu said. "But its advent proves that the domestic scientific research team has the ability to overcome important technical challenges on the way to developing ChatGPT-like products."

图源：视觉中国

“大型语言模型几乎是被国外垄断的。他们只开发API（应用程序编程接口），或者干脆不向我们开放。”

"Large-scale language models are almost monopolized by foreign countries. They only develop APIs (application programming interfaces) or do not open them to us," Qiu said.

“我们希望与大家分享MOSS及其模型代码和开发经验，并希望中国能够在大规模语言模型领域走在世界前列。”

"We want to share MOSS and its model codes and development experiences with everyone, and hope that China can be at the forefront of the world in terms of large-scale language models."

MOSS可执行对话生成、编程、事实问答等一系列任务，打通了让生成式语言模型理解人类意图并具有对话能力的全部技术路径。这条路径的走通，为国内学术界和产业界提供了重要经验，将助力大语言模型的进一步探索和应用。

邱锡鹏说，“期待复旦团队和上海人工智能实验室继续合作，通过MOSS以及后续的研究探索，推动AI普惠，尽快赋能国内的人工智能产业。”

"We look forward to the continued cooperation between the Fudan team and the Shanghai laboratory, through MOSS and subsequent research and exploration, to promote AI inclusiveness and empower the domestic AI industry as soon as possible," he added.

对于国内自主研发的MOSS，你有什么看法呢？

记者：林淑娟

编辑：左卓

实习生：李龙一

来源：澎湃新闻 上观新闻 IT之家 中国日报