【访谈】王耀南院士：人工智能如何更好推动数字经济发展

人工智能产业未来发展趋势如何？目前面临着怎样的挑战？机器人如何与人工智能有机结合？围绕相关问题，对话中国工程院院士、中国数字经济50人论坛委员、机器人技术与智能控制专家王耀南。

Q1：作为最热门的前沿技术之一，您认为人工智能将如何推动数字经济高质量发展？

王耀南院士：首先，要将人工智能领域的几个关键技术应用于经济产业，一是人工智能的深度学习；二是人工智能的的感知功能；三是人工智能的记忆推理；四是人工智能的联想，包括人工智能的分析判断以及人工智能的智能控制。

人工智能是一种赋能技术，赋能产业的数字化、网络化和智能化转型升级。例如将人工智能应用到智能制造领域，可以通过人工智能的感知技术、认知技术和智能控制技术，来控制机器人完成自动化的生产和智能化的转型。

在农业领域，可以应用人工智能技术对育种进行分析。在交通领域，可以对车流量进行监控以解决交通堵塞问题。此外，在智慧医疗、智慧政务及教育等领域都可以应用人工智能技术。 CXO UNION-CXO联盟（cxounion.cn）

Q2：今年以来，人工智能大模型发展迅速，市场上相关产品层出不穷，您认为未来相关产品发展有何趋势？产品如何在竞争中取得优势？

王耀南院士：随着ChatGPT的迅速发展，国内企业和研究机构快速跟进人工智能大模型产品的研发，因此我认为要清醒的认识、清醒的应用，并清醒的找准自己的特点。

第一，要重视人工智能大模型，并且找准应用场景。大模型并非万能，关键是要以问题和需求为导向，来构建人工智能大模型，以此解决我们的痛点和难点问题。例如针对医疗行业，就要构建医疗大模型，才能更加专业并且更加精准的进行治疗。

而制造业的需求与医疗行业不同，强调“高速、高精、高实时”的制造，就要有针对制造业的轻量化模型，因此要根据不同的应用场景研发具有针对性的大模型，进而解决不同的问题。

Q3：您认为机器人如何借助人工智能发展更加智能化？人工智能大模型将对机器人行业发展带来哪些机会？

王耀南院士：在今年的世界机器人大会上，我以《集群机器人协同制造技术应用及发展趋势》为主题作报告，其中我认为集群机器人可以与人工智能大模型相结合。因为机器人是一个特别大的系统，需要用到人工智能大模型通过深度强化学习，来解决多目标、多机器人的优化调度。 CXO UNION-CXO联盟（cxounion.cn）

此外，如果要提升机器人的自主性和智能化水平，就要让机器人能够回答人类很多问题，在解决很多人类难以执行的任务时，也可以与大模型有机的结合起来。其中，机器人要做好执行规划，另外大模型要做好决策分析和判断，进而提升机器人的认知能力，这是机器人产业未来发展中，非常重要的方向。

Q4：智能产业发展目前面临哪些困难？对此，您有哪些建议？

王耀南院士：我个人认为,人工智能产业主要分为两类，一是智能产业本身开发各种智能产品，例如开发无人驾驶电动汽车可以解决能源问题，其中涵盖了能源、通讯、导航、地图创建、互联互通以及车载安全控制等，因此这是综合性非常强的产业。 CXO UNION-CXO联盟（cxounion.cn）

二是人工智能赋能到其他行业，例如农业、工业和医疗等，进而形成庞大的产业。因此人工智能可以向上述两个方向发展。

此外，目前人工智能发展过程中所面临的困难和挑战非常多，第一，在人工智能的模型、训练的算法以及核心硬件等方面存在痛点。因为在计算的过程中消耗能量较高，所以要找出怎样在提高算力的前提下降低能耗的方案。而如何开发出计算效率高的计算架构芯片，是全世界所面临的挑战性问题。

第二，要解决现有人工智能模型的可解释性、安全性和可靠性，以及要让大模型在遇到各种复杂干扰时，仍能保证稳定运行。这是一个需要大家继续努力，攻坚克难的问题。

第三，要解决数据安全性问题，还需要在数据的挖掘、数据分析、数据安全隐私以及数据提取方面找出解决方案，要使用有用的数据来训练网络。 CXO UNION-CXO联盟（cxounion.cn）

Q5：对于人工智能产业人才的培养，您有哪些建议？

王耀南院士：既然行业面临如此多的挑战，那么就一定要培养好人工智能人才。近几年各高校陆续开放人工智能专业，培养人工智能人才。但人工智能的算力人才、算法人才、数据人才以及应用场景人才都要分类培养。也就是需要从基础研究型、技术工程师、应用场景三个维度进行培养。

翻译：

Academician Wang Yaonan: How can artificial intelligence better promote the development of digital economy

What is the future development trend of artificial intelligence industry? What are the current challenges? How to combine robots with artificial intelligence? Around relevant issues, talk to Wang Yaonan, academician of the Chinese Academy of Engineering, member of the China Digital Economy 50 Forum, robot technology and intelligent control expert. CXO UNION-CXO联盟（cxounion.cn）

Q1: As one of the hottest cutting-edge technologies, how do you think artificial intelligence will promote the high-quality development of the digital economy?

Wang Yaonan: First of all, several key technologies in the field of artificial intelligence should be applied to the economic industry, one is the deep learning of artificial intelligence; The second is the perception function of artificial intelligence; Third, the memory reasoning of artificial intelligence; The fourth is the association of artificial intelligence, including the analysis of artificial intelligence and the intelligent control of artificial intelligence.

Artificial intelligence is an enabling technology that enables the digitalization, networking and intelligent transformation and upgrading of the industry. For example, applying artificial intelligence to the field of intelligent manufacturing can control robots to complete automated production and intelligent transformation through the perception technology, cognitive technology and intelligent control technology of artificial intelligence.

In the field of agriculture, artificial intelligence technology can be applied to analyze breeding. In the field of transportation, traffic flow can be monitored to solve the problem of traffic congestion. In addition, artificial intelligence technology can be applied in areas such as smart healthcare, smart government and education. CXO UNION-CXO联盟（cxounion.cn）

Q2: Since the beginning of this year, AI large models have developed rapidly, and related products have emerged in an endless stream on the market. What do you think are the trends of related product development in the future? How does the product gain an advantage over the competition?

Academician Wang Yaonan: With the rapid development of ChatGPT, domestic enterprises and research institutions have quickly followed up the research and development of AI large model products, so I think we should have a clear understanding and clear application, and be clear about their own characteristics.

First, we should pay attention to the large model of artificial intelligence and find the correct application scenarios. The big model is not omnipotence, the key is to build a big model of artificial intelligence based on problems and needs, so as to solve our pain points and difficult problems. For example, for the medical industry, it is necessary to build a medical model in order to be more professional and more accurate treatment. CXO UNION-CXO联盟（cxounion.cn）

The demand of the manufacturing industry is different from that of the medical industry, emphasizing the “high-speed, high-precision, high-real-time” manufacturing, it is necessary to have a lightweight model for the manufacturing industry, so it is necessary to develop a targeted large model according to different application scenarios, and then solve different problems.

Q3: How do you think robots can become smarter with the help of artificial intelligence? What opportunities will the AI Grand model bring to the development of the robotics industry?

Academician Wang Yaonan: At this year’s World Robot Conference, I gave a report on the theme of “Cluster robot collaborative Manufacturing Technology Application and development Trend”, in which I believe that cluster robots can be combined with large-scale artificial intelligence models. Because the robot is a particularly large system, it is necessary to use a large artificial intelligence model to solve the optimization of multi-objective and multi-robot scheduling through deep reinforcement learning.

In addition, if we want to improve the autonomy and intelligence level of robots, it is necessary to make robots able to answer many questions of humans, and can also organically combine with large models when solving many tasks that are difficult for humans to perform. Among them, the robot should do a good job of execution planning, and the large model should do a good job of decision analysis and judgment, so as to improve the cognitive ability of the robot, which is a very important direction in the future development of the robot industry. CXO UNION-CXO联盟（cxounion.cn）

Q4: What are the difficulties facing the development of intelligent industry? What are your suggestions?

Academician Wang Yaonan: I personally believe that the artificial intelligence industry is mainly divided into two categories. One is the intelligent industry itself to develop a variety of intelligent products, such as the development of driverless electric vehicles can solve the energy problem, which covers energy, communication, navigation, map creation, interconnection and vehicle security control. So this is a very comprehensive industry.

The second is that artificial intelligence enables other industries, such as agriculture, industry and healthcare, to form huge industries. So AI can develop in both directions.

In addition, there are many difficulties and challenges in the current development process of artificial intelligence. First, there are pain points in the model of artificial intelligence, the training algorithm and the core hardware. Because of the high energy consumption in the calculation process. It is necessary to find out how to reduce energy consumption under the premise of increasing computing power. How to develop a computing architecture chip with high computational efficiency is a challenging problem facing the world. CXO UNION-CXO联盟（cxounion.cn）

Second, it is necessary to address the interpretability, security and reliability of existing AI models. As well as to ensure stable operation of large models in the face of various complex disturbances. This is a problem that requires all of us to continue to work hard to overcome.

Third, to solve the problem of data security, it is also necessary to find solutions in data mining, data analysis, data security and privacy, and data extraction, and to use useful data to train the network.

Q5: What are your suggestions for the cultivation of talents in the artificial intelligence industry?

Academician Wang Yaonan: Since the industry is facing so many challenges, it is necessary to cultivate artificial intelligence talents. In recent years, universities have opened artificial intelligence majors to train artificial intelligence talents. However, the computing power talents, algorithm talents, data talents and application scenario talents of artificial intelligence should be classified and cultivated. In other words, it is necessary to train from three dimensions: basic research type, technical engineer and application scenario. CXO UNION-CXO联盟（cxounion.cn）

由CXO UNION-CXO联盟（cxounion.cn）转载而成，来源于财联社；编辑/翻译：CXO UNIONCXO联盟小U。

如需加入CXO UNION（CXO联盟）高管社群，请联系社群小伙伴哦~

免责声明: 本网站(http://www.cxounion.cn/)内容主要来自原创、合作媒体供稿和第三方投稿，凡在本网站出现的信息，均仅供参考。本网站将尽力确保所提供信息的准确性及可靠性，但不保证有关资料的准确性及可靠性，读者在使用前请进一步核实，并对任何自主决定的行为负责。本网站对有关资料所引致的错误、不确或遗漏，概不负任何法律责任。

本网站刊载的所有内容(包括但不仅限文字、图片、LOGO、音频、视频、软件、程序等) 版权归原作者所有。任何单位或个人认为本网站中的内容可能涉嫌侵犯其知识产权或存在不实内容时，请及时通知本站，予以删除。