人工知能 上手に使って住み良い社会に

The Yomiuri Shimbun
Artificial intelligence: Use it well for a society where humans can thrive
人工知能　上手に使って住み良い社会に

The recent feat achieved by AlphaGo was a marvel of striking progress in artificial intelligence (AI) technology.
AlphaGo, an AI-based computer program developed by a British corporation under the umbrella of Google Inc. of the United States, has won against the world’s top Go player, South Korea’s Lee Se Dol, 4-1.
　人工知能の技術が目覚ましい進化を遂げていることに驚かされた。
　米グーグル傘下の英企業が開発した人工知能の「アルファ碁」が、韓国棋院の世界トップ棋士、李世●九段に４勝１敗で勝ち越した。（●＝「石」の下に「乙」）

Previously, AI programs had defeated skilled human players in the fields of chess and shogi. However, it was said that it would take 10 years to see an AI system win against human players in the world of Go.
　チェスや将棋では、人工知能が実力者を破っている。だが、囲碁で勝つには、この先１０年を要すると言われていた。

It was cited as a high hurdle that the surface of a Go board is broad, and that there are an immeasurable number of choices for moves to be made in playing a match.
盤面が広く、着手の選択肢が桁違いに多いことが、高いハードルとされた。

No wonder Japan’s top master of Go, Yuta Iyama, described AlphaGo’s victory as “shocking.”
　日本の第一人者である井山裕太棋聖が「衝撃的」と語ったのも、無理はない。

AlphaGo is able to learn on its own which are the best moves to make in playing a game. The program has read about 30 million images depicting the stages of development in games played by professional and other players. Then it has learned what kinds of patterns are preferred at which stages of development in a game, giving itself a better chance of winning. This has made AlphaGo ever more highly skilled in playing a match.
　アルファ碁は、より良い一手を自ら学ぶ能力を備えている。プロらによる対局の盤面の画像を３０００万枚も読み込み、どのようなパターンになれば勝ちやすいかを学習し、棋力を高めた。

That approach, called “deep learning,” is one of the state-of-the-art AI theories. To put its victory in a human context, AlphaGo has developed “taikyokukan” (broad-based perspective).
　「ディープ・ラーニング（深層学習）」という最新理論だ。人間で言えば、「大局観」を養ったことが勝因だろう。

AI technology has already been put to practical use, such as in image searches. It may be possible that the spread of more advanced AI technology could greatly change our day-to-day life. There also are concerns that the technology could take over clerical work carried out by humans.
　人工知能は、画像検索などで既に実用化されている。より高度な人工知能が普及することで、日常生活が大きく変化する可能性がある。事務などの仕事が取って代わられるといった懸念もある。

It is necessary to live with AI technology and make good use of it.
　人工知能と共存し、活用していくことが求められる。

Myriad potential benefits

The self-driving car system is designed to prevent accidents by perceiving vehicles and pedestrians around an autonomous car and predicting how they would move.
　車の自動運転では、周辺の車両や歩行者を察知し、動きを予測することで、事故を防ぐ。

The use of diagnostic imaging technology will increase the probability of detecting cancer cells. It will also make it possible to predict how the condition of a cancer patient would develop, thereby enabling the patient to receive appropriate treatment.
　画像診断では、がん細胞を発見する確率が高まろう。病状の進行を予測して、適切な治療に結び付けることも可能になる。

Progress is being made in developing the technology to instantly distribute to smartphones information about which route should be taken in evacuating at the time of tsunami and other disasters.
　津波などの災害時に、どのルートで避難すれば安全か、という情報を瞬時にスマートフォンへ配信するための技術開発が進む。

We hope the public and private sectors will join hands in fully tapping into the advantages of AI technology.
　官民が連携し、人工知能のメリットを最大限に引き出したい。

In the United States, Google, IBM and others are fiercely competing with each other in applying AI technology to the analysis of big data, the development of pharmaceutical products and for other purposes.
　米国では、グーグルやＩＢＭなどが、ビッグデータの解析や医薬品開発などへの応用で激しい競争を繰り広げている。

It is disturbing to see that the presence of Japanese corporations is weak in this respect. Because of budgetary constraints and other reasons, these companies have hardly been able to apply their research results to the commercialization of products. Research papers originating in Japan account for a slumping 2 percent to 3 percent of those submitted from around the world.
　これに対し、国内企業の存在感が薄いのが気がかりだ。予算の制約などで、研究成果が、ほとんど製品化につながっていない。世界の中で日本発の論文が占める割合は、２～３％と低迷している。

It is necessary to quickly nurture young researchers and engineers in this country.
　若手の研究者や技術者の育成を急ぐ必要がある。

The main pillar of the 5th Science and Technology Basic Plan, finalized by the government at a Cabinet meeting in January, is to realize “a supersmart society” through the use of AI technology.
　政府は、１月に閣議決定した第５期科学技術基本計画の柱に、人工知能による「超スマート社会」の実現を据えている。

If robotics technology, Japan’s field of expertise, is combined with AI, a mix of these technologies could be expected to make up for labor shortages in a society whose population is decreasing.
　日本の得意分野であるロボット技術に人工知能を組み合わせれば、人口減社会での労働力不足を補う役割も期待できるだろう。

(From The Yomiuri Shimbun, March 16, 2016)