Taiwan’s future in AI development

By Kung Ming-hsin 龔明鑫  / 

Thu, Oct 05, 2017 - Page 8

The Taipei Times published an opinion piece by Digital Intelligence Service Alliance chairman Robert Wang (王定愷), in which he wrote that the Cabinet’s policies on artificial intelligence (AI) need to venture beyond a “scientific research” mindset (“AI development needs strategic shift,” Sept. 23, page 8).

Wang wrote that oversight and guidance for AI policy should be taken to a higher level so that it can be implemented by several ministries acting in unison. Wang’s article highlights the need for a clear explanation of the government’s current AI policies.

Speaking at the opening of the World Conference on Information Technology in Taipei on Sept. 11, President Tsai Ing-wen (蔡英文) said that, with Taiwan having entered the digital age, the government has launched the “Asian Silicon Valley” development plan.

It has also drawn up a special budget to speed up strengthening Taiwan’s digital infrastructure. Tsai said she believes that Taiwan will be able to emerge from the last industrial age to become a model nation in the digital economy.

The next day, Premier William Lai (賴清德) presided over a Cabinet meeting for the first time. After listening to a progress report on the “Asian Silicon Valley” initiative, Lai highlighted three main tasks: strengthening infrastructure, promoting digital education and relaxing legal restrictions.

He said that improved infrastructure would allow Taiwanese to benefit from progress in smart technology and improve their quality of life.

Tsai and Lai not only reiterated the government’s determination to develop a digital economy, but also its resolve not to restrict itself to a research-oriented mindset. The government will address comprehensive concerns, such as infrastructure, talent, laws and regulations, and recognize the importance of application services.

Breakthroughs in AI technology will hasten the arrival of the Internet of Everything (IoE). Deep learning technology will greatly improve the power of machine vision, along with facial, voice and image detection and recognition, opening up AI applications in diverse industries, such as “smart” manufacturing, “smart” medicine and driverless cars.

Taiwan’s best opportunities in AI development can be found, first, in AI on the edge and, second, in the cloud.

AI on the edge involves devices that can connect to the cloud. In the age of PCs, a computer’s functions were based on its hardware and peripherals, but in the mobile age, most functions are located in the cloud, with the role of the connected hardware generally being reduced to a minimum.

In the IoE, it will no longer be possible to locate all functions in the cloud. No amount of bandwidth will be able to support the full burden of the IoE, so demands are sure to arise for AI and data security on the edge — this presents opportunities that match Taiwan’s strengths.

The cloud and cloud applications involve data, which lies at the heart of both the digital economy and AI. Taiwan has many comprehensive big databases that are the envy of the world, including an accurate and detailed national health database, comprehensive manufacturing and performance data for all industrial sectors, and agricultural and soil data gathered over several decades.

AI development requires data that is comprehensive and analyzable. It also requires innovative algorithms and cooperation among specialist fields to develop applications. Much work remains if these things are to be achieved.

First, there is the problem of how to link data across departments as quickly as possible while protecting personal data. This includes data in such realms as health insurance, labor, education and finance. The more integration there is, the greater the likelihood of achieving innovation.

Second, algorithms are essential for innovation. Besides nurturing and strengthening AI and software talent, it will be necessary to boost basic hardware computing power.

Finally, complete business ecosystems must be constructed for different specialist sectors — this is an important goal of the “Asian Silicon Valley” initiative.

The government has ministries and ministers without portfolio working on these tasks at an interministerial level.

However, besides the central government, manufacturing ecosystems include other participants, such as local governments, academia, research institutes, big corporations, small and medium-sized enterprises, and start-ups.

All these sectors and more will hopefully take part in this process and share in the opportunities that AI will bring.

Kung Ming-hsin is the deputy minister of economic affairs and cochief executive of the Asia Silicon Valley Development Agency.

Translated by Julian Clegg