Guide:
Middle school students in prestigious schools in Beijing can develop AI models and even publish papers based on their interests, while principals in remote areas are still hesitating about "artificial intelligence education, will the high school entrance examination pass the exam? Will the college entrance examination give extra points?"
The educational gap has been around for a long time, and it cannot be reversed only by some enterprises or educational institutions, but we can at least start with artificial intelligence education.
"Some children are still training after class, and they are reluctant to leave. Sometimes it is 10 o'clock in the evening, and the children don't want to go home. I have led students to robot competitions for many years. Good courses can play a role in stimulating students' interest. Very positive."
At a public event recently, Mr. Yuan Zhongguo, head of the information technology teaching and research group of the High School Affiliated to Renmin University of China, shared his feelings about an artificial intelligence course in the school. This course uses an important tool - the self-driving car Amazon DeepRacer from Amazon cloud technology.
However, don’t think that only students from prestigious schools in first-tier cities have the opportunity to get in touch with the latest artificial intelligence education. In order to provide more young people with the opportunity to learn and practice artificial intelligence and lower the threshold of artificial intelligence education, on March 4, Amazon Cloud Technology announced a strategic cooperation with the Science and Technology Innovation Education Collaboration of the Chinese Society of Education and the Shanghai Artificial Intelligence Industry Association. Amazon DeepRacer training will be provided to 100,000 Chinese teenagers in the next three years.
In recent years, AI education projects for teenagers have emerged one after another. What is the magic of Amazon DeepRacer? How did it spark interest in artificial intelligence among teens?
Ke Ruifeng (left), Secretary-General of the Science and Technology Innovation Education Collaboration of the Chinese Society of Education, Gu Fan (middle), General Manager of the Strategic Business Development Department of Amazon Cloud Technology Greater China, and Zhong Junhao (right), Secretary-General of the Shanghai Artificial Intelligence Industry Association, attended the ceremony Strategic Cooperation Launching Ceremony
What's so amazing about Amazon DeepRacer?
Amazon DeepRacer is a 1/18 scale fully automatic driving racing car launched by Amazon Cloud Technology in 2018. Its biggest charm is that even novice users without artificial intelligence expertise can quickly get started training with just a few parameters. Model, learn artificial intelligence through the form of physical car competition.
“Amazon DeepRacer门槛低、趣味性强、可操作性高。我们希望利用这一独特的学习工具，开发系统的课程并推出丰富的教育和培训活动，降低大众学习人工智能的门槛，激发他们对机器学习的兴趣与热情，助力各阶段人工智能人才培养。” 亚马逊全球副总裁、亚马逊云科技大中华区执行董事张文翊表示。
亚马逊云科技大中华区战略业务发展部总经理顾凡认为，Amazon DeepRacer赛车虽然小巧，但它不是玩具，而是一个可以在云上训练的低门槛的AI学习工具。参加Amazon DeepRacer比赛，就可以体验算法工程师在现实中开发模型、解决实际问题的过程。具体来说，Amazon DeepRacer与真实的AI开发有三个一致：
原理一致：Amazon DeepRacer的原理是强化学习，即通过设置正确的奖励函数，通过持续给算法反馈使其快速迭代，赢得比赛。
过程一致：Amazon DeepRacer可以让每个参与者亲身体验AI算法训练的全过程，包括数据准备、训练模型、模型部署、参数调优、重新迭代。
体验一致：要想训练出好的模型，需要无数次试错，不断优化模型。这与真实世界中AI算法开发体验是一致的。通过Amazon DeepRacer，青少年可以深刻体会到，成功的AI模型就是从海量的试错中走出来的。
AI是专才教育还是通识教育？
随着第三次人工智能浪潮的兴起，AI产业对人才的需求井喷。根据2020年人社部的调查，我国人工智能人才缺口达到500万，供求比例为1:10。如何解决人才短缺这一AI产业发展瓶颈，人工智能从专才教育转变为通识教育是必经之路。
“如今人工智能正在赋能百业。我的理解，人工智能是未来的通识教育，不是专业教育。我们需要从青少年抓起，培养他们的人工智能素养。”上海市人工智能行业协会秘书长钟俊浩表示。
当然，AI要从专才教育转变为通识教育，还面临很多挑战。比如师资水平、课程设置、东西部教育资源不平衡等。
比如，以人大附中为代表的一线城市名校中，一些优秀的中学生已经可以结合古诗词、RAP等个性兴趣开发模型，有的还发表了论文。亚马逊云科技还将真实行业案例进行简化、提炼，融入课程，支持青少年用AI解决实际问题。比如，基于亚马逊云科技某能源行业解决方案，人大附中的同学们建立了一个时序预测模型，能够基于某地过往温度数据预测未来几天的温度。
与之形成鲜明对比的是，中西部和偏远地区在人工智能教育方面还存在不小差距。
“有些偏远地区的老师听课之后感慨，说大城市孩子开始学的这些内容，他们老师都还不懂。有的地区校长的理念还没转变过来，会问人工智能教育中考考吗？上清华、北大能加分吗？有的老师学完了回去想开课，领导不支持，没有课时。” 袁中果老师表示。
为了缩小教育鸿沟，让更多青少年受益于更普惠的AI教育，亚马逊云科技与中国教育学会科创教育协作体及上海市人工智能行业协会三方计划在课程研发、师资培训、平台支持和以赛促学四个方面加大投入。比如，共同开发更适合中国青少年的课程，通过定期组织名师交流会进行师资培训，提供长期免费的Amazon DeepRacer学习平台和云资源等。
“我们从2020年就开始举办Amazon DeepRacer相关活动，如今已经有400所学校的上万名学生学习了相关课程。今年最大的不同就是要把活动从过去的一、二线城市的重点学校、国际学校，下沉到更多地区和院校。” 顾凡表示。
人工智能教育无法用“昨天的课程，今天的老师，培养出未来的人才。” 如何降低青少年学习人工智能的门槛，构建更普惠的AI人才培养模式，值得每一家科技企业与教育机构思考。
END
本文为「智能进化论」原创作品。