“AlphaGo Zero” and The Power of Learning
– Musings of Dr. Jamie C. Hsu, 11.7.2017
The Google DeepMind program AlphaGo Zero recently achieved superhuman level and beat the best GO players in the world — without any human help!
The system starts with a neural net that does not know anything about GO. A set “rules of the game” is prescribed, and then the model generates data and learns on its own with Deep Learning technology. It has been able to:
- Beat the previous version of AlphaGo (Final score: 100–0).
- Learn to perform this task from scratch, without learning from previous human knowledge or recorded game plays.
- Attain World champion level GO playing in just 3 days of training.
This accomplishment has clearly demonstrated the superiority of smart algorithms and math over big data and memory. While data is obviously valuable in many cases, the breakthrough of AlphaGo Zero clearly shows us the Power of Learning.
The same can be said about an individual or organization. You can improve and perform based on a vast amount of data, memory bank, and analytical tools. But that will take a lot of effort, time, and money. You may even drown in the sea of information and data. The brute force approach is used by many schools in educating children, and in many organizations that collect vast amounts of big data.
Although this diligent effort will undoubtedly produce some results eventually, a better way is to teach the children and the organizations critical skills and a mindset of “how to learn.” With the right learning methods, environment, and mindset, people and organizations can grow and improve with minimal intervention. In the past, many books have been written about Learning Persons and Learning Organizations. Maybe it is time to dust them off the bookshelf and take another look.
We should be inspired, not intimidated, by the accomplishments of AlphaGo Zero and the ever-growing capability of artificial intelligence in our daily lives.
Let’s become a Learning Person and build a Learning Organization.
“AlphaGo Zero” – 學習的力量
-作者 許俊宸博士
-中譯 薛乃綺
最近Google DeepMind計畫所開發的人工智慧AlphaGo Zero，在無需要任何人類的協助下，擊敗全世界最強的圍棋高手，達到超水準境界！AlphaGo Zero從一個完全不知道圍棋規則的神經網絡開始；”遊戲規則”制定後，讓模型生成數據，並利用深度學習技術自行學習。它已經能夠：
- 以100:0擊敗之前的AlphaGo版本。
- 學習從頭開始執行任務，無須透過學習之前人類的知識或錄製遊戲。
- 僅用3天訓練時間就達到世界圍棋冠軍等級的水準。
這項成就清楚地揭露，智能演算法比大數據與內存記憶體更加重要。儘管數據在許多情況下還是顯著的價值，但AlphaGo Zero的突破清楚地向我們展示學習的重要与力量。
這對於個人或組織也是如此。你可以根據大量的數據、內存記憶庫及分析工具進行改善與執行。但這需要花費很多的精力、時間和金錢；而且你甚至可能會被過量的訊息及數據給淹沒。許多學校都在用強攻的方式來教育孩子、也同樣也許多組織不停在大量地收集資訊。
雖然這樣勤奮的努力最終會產生一些成果，但更好的方式也許是培養孩子或企業”如何學習”這種的關鍵技能和思維方式。有了正確的學習方法、環境和思維方式，人員和企業組織就能在最少干預下達到改善與成長的目標。過去有很多關於學習者和學習型組織的書籍。也許是時候再將它們從書架上再拿下來看看了！
看待AlphaGO Zero的成就以及不斷在我們日常生活中茁壯成長的人工智能，我們應該受到啟發、而不是被它們嚇倒。
讓我們成為一個學習者、建立一個學習型的組織吧！
