Figure source | Zoom live images
Finishing & Zebian | yugao
Exhibition | CSDN (ID: CSDNnews)
Microsoft worked after 13 years, in November 2019, Microsoft executive vice president, Artificial Intelligence and Research Division Leader Harry Shum announced that it will officially leave 1 January 2020, but it was not announced that he will whereabouts where. March 5, appeared to Yang Shen Qinghua University, in this special period, through live video, Tsinghua University Qiu Yong presented letters of appointment to sink to the ocean, was officially re-appointment of Harry Shum Institute for Advanced Study, Tsinghua University professor appointed double. Institute for Advanced Study in the official website of Tsinghua University, has been able to found doctoral enrollment plan to Shen Yang.
The reason why is "the reappointment," because this is the second time Shum was appointed as professor of Tsinghua University, 15 years ago, Tsinghua University held the first appointment ceremony to sink to the ocean. After going around in circles, groping for 15 years, he was again stationed in this old familiar institution.
March 5, live platform by Zoom, Shen was the theme of the Ocean Engineering Responsible AI course lectures.
In his speech, he talked about the sink for some current views and research of AI, such as AI and AI prejudice the interpretability issues to focus on the ocean.
FIG. Source: Live screenshots
He said, to design and build a responsible AI is very important, which is one of the challenges we face AI interpretability. While the current AI applications have been around everywhere, but the lack of AI interpretability, our lack of understanding of AI decision making mechanism, the model can explain the complexity is very high, even if the model can predict a higher accuracy rate, but we do not I know why. Therefore, interpretable model is very important.
FIG. Source: Live screenshots
AI bias is another important challenge facing AI research and applications. Shen cited two examples to the ocean, to explain the bias from sample data used in training, and the elimination of prejudices embedded in the text. He says, prejudices cause various problems, the key is that even the experts, but also know there is a problem, but do not know where the problem is, and the cause of the problem, let alone solve the problem.
沈向洋总结，我们是第一代和 AI 共存的人类，无论喜欢与否都别无选择，但是我们能决定该用何种方式来构建 AI 以及使用AI，并以一个直击心灵的问句结束了演讲：“我们能接受一个 AI 做的决定、我们人类无法理解的世界吗？”
最后，我们来简单回顾下沈向洋的从业经历。无论是在学术界还是业界，沈向洋的职业和研究生涯可谓成绩卓越。
自博士毕业后，沈向洋经由陆奇引荐，于 1996 年 11 月加入美国微软研究院，开始了他在微软 23 年的职业生涯。
1998 年，设在北京的微软中国研究院（Microsoft Research China，后改名为微软亚洲研究院）成立时，他应院长李开复的要求前往加盟，并于 2004 年担任院长、首席科学家。
2019 年 11 月 14 日，沈向洋宣布将离开微软，但还会作为萨蒂亚·纳德拉的个人顾问。
他曾是硅谷大科技公司中职位最高的华人高管，不仅负责推动微软整体 AI 战略及涵盖基础架构、服务、应用程序和代理的前瞻性研发工作，还领导了包括 Bing 在内的以 AI 为重点的产品组，以及微软研究院及其与公司整个工程团队的集成。
沈向洋被称为微软的中国先生，也被视为中国的微软大使，他曾两度在世界人工智能大会上代表微软发表演讲。
沈向洋主要专注于计算机视觉、图形学、人机交互、统计学习、模式识别和机器人等方向的研究工作，是计算机视觉和图形学研究的世界级专家，先后发表关于计算机视觉、计算机图形学、图形识别、统计学习和机器人科学方面的论文 100 多篇，拥有美国专利超过数十项。
沈向洋前后获得香港大学电机电子工程系硕士学位和卡内基·梅隆大学计算机学院机器人专业博士学位，而在博士时期，沈向洋就已经具有了对行业发展做出了极有价值的前瞻，博士论文是世界上最早有关由照片转换成虚拟现实的研究，所设计的四分树样条数函数算法，是世界上最好的运动参数估计算法之一。
在计算机视觉领域，沈向洋于 1994 年公布第一个三维全真模型时，几乎还没有人在这一领域进行类似的工作；1996 年发明同心拼图法，将四维全光函数简化为三维函数，大幅降低需采集的数据量。
这样一位世界级专家任教于清华，相信对清华乃至中国 AI 的发展都会产生积极的影响。
推荐阅读
☞华为补助武汉员工，最高每日 2000 元；iPhone SE 2 量产或推迟；PowerShell 7.0 发布 | 极客头条
☞腾讯云“抢救”微盟！开 766 次在线会议、调拨 100 多台服务器、闹钟只敢定 2 小时
☞6 个步骤，教你在Ubuntu虚拟机环境下，用Docker自带的DNS配置Hadoop | 附代码
☞开发项目时如何选择区块链平台？我们分析了以太坊、Bitcoin via RSK、Ardor三个有趣的平台来给你回答！
你点的每一个在看，我认真当成了喜欢
