人機大戰，巔峰競技 - Game V: AlphaGo笑到最後
After suffering its first defeat in the Google DeepMind Challenge Match on Sunday, the Go-playing AI AlphaGo had beaten world-class player Lee Sedol for a fourth time to win the five-game series 4-1 overall. The final game proved to be a close one, with both sides fighting hard and going deep into overtime.
The win came after a "bad mistake" made early in the game, according to DeepMind founder Demis Hassabis, leaving AlphaGo "trying hard to claw it back." By winning the final game despite its blip in the fourth, AlphaGo has demonstrated beyond doubt its superiority over one of the world's best Go players, reaffirming a major milestone for AI in the process.
It was "the most mind-blowing game experience we've had so far," said Demis Hassabis at the post-match press conference, with an "incredibly close and tense finish." Lee said that he felt sorry the match was coming to an end, while expressing how difficult it has been from a psychological perspective.
Both played strongly and the final game was by far the best in the match. Google's AI program actually made a serious "blind spot" error similar to the one that cost it the fourth game. However, it quickly collected itself and played nearly error-free after that. Sedol also played well, but AlphaGo chipped away at his lead and took the match after it went into tense "endgame" state, ended with Sedol’s resignation.
DeepMind's AlphaGo program stunned the Go-playing world by beating 18-time world champion Lee Sedol … thanks to its advanced system based on deep neural networks and machine learning. Lee was competing for a $1 million prize put up by Google, but DeepMind's victory means it will be donated to charity.
After the match, South Korea's Go Association awarded AlphaGo the highest Go grandmaster rank – ninth dan. It was given in recognition of AlphaGo's "sincere efforts" to master Go's Taoist foundations and reach a level "close to the territory of divinity".
References
Google's AlphaGo AI beats Lee Se-dol again to win Go series 4-1
By Sam Byford on March 15, 2016 05:00 am
http://www.theverge.com/2016/3/15/11213518/alphago-deepmind-go-match-5-result
建中資優生在各種棋類及橋藝都有不凡表現。
因較少下圍棋，對台灣圍棋之碁壇動態不太清楚。〈較常打橋牌、下象棋或撞球～都不精。〉
唯據知有的業餘好手其實已具職業相當水準，只是因故不熱衷應試取得職業棋士段位證照。
例如業餘六段棋士沈君山前輩；業餘六段棋士黃士傑先生。〈福到格主也是深藏不露好手。〉
參考「台灣圍棋發展協會資訊」：
福到格主是圍棋業餘好手；據悉還贏過咱們台灣職業棋士。
〈印象裡，福到格主也是UDN界的象棋好手。〉
如金大俠所言，台灣的職業圍棋制度確實不完善。
有的職業棋士狀況不穩定，因環境等故無法維持一定水準。
關於圍棋史實，福到兄了然於胸，佩服。下文係部分史實。
參閱《論圍棋佈局理論及思想之演變》：吳昌政
这是我对第四局的评论：
I think, this game, Lee is simply lucky.
Alpha-Go can be improved even more in the future.
去棋社觀戰
棋友慫恿我與某職業棋士對弈 當然是無彩不成棋(他剛下完授子棋贏了錢)
他根本不認識我 照說應該是先問局差才對
結果 我猜到白子 更要命的是我贏棋
當然 此君我必須說棋品有問題(對業餘怎能下分先呢)
金大俠 於 2016/03/21 10:52回覆
戰前, 專家九段對初段手合是四子 戰後為三子
對四五段都是兩子
吳清源巔峰時 有人吹捧為11段的棋力 理由是對藤澤的戰績超好
但坂田自六番賽勝吳後 吳已逐漸喪失霸業
六強賽後對坂田更是輸多贏少 內容亦不復當年的質
1955 工藤紀夫以初段 持白勝影山利郎五段 轟動一時
1960年代後 初二段持白勝八九段 已非新聞
這種例子在1980後的韓國更是屢見不鮮
遂有延續傳統晉升制度與大比賽獲頭銜者逕升九段之作法
後 大陸亦跟進
順便一提 1980初期日本仍是幾百年來的圍棋王國
但現不當大哥已三十年了 頂尖棋士遇上大陸 韓國高手
想開胡實屬不易
UDN中懂棋者實在不多矣～UDN中諳象棋者不少；諳圍棋者較少或深藏不露。
大俠保守估計，目前的AlphaGo或有11段～金大俠肯定是AlphaGo大粉絲，致敬。
有專家如是說：AlphaGo在技術面至少有職業七段〈七段至九段均屬高段棋士〉，在整體戰力〈技術＋穩定性＋持久耐力〉則有九段大師級功力。聊供參考。
至於AlphaGo盤面上贏太多時會下緩手慢棋，盤面上落後時則會急起力追～亦有諳圍棋之工程師看法是：AlphaGo「遇強則強，遇弱則弱」，算是無傷大雅之Bug。
當然，AlphaGo既有自學能力，也可能具有人性～未來人工智慧機器人發展將不可限量，人類失業潮亦將剉得等矣。
可惜，人工智慧電腦先後戰勝了西洋棋王及圍棋大師，但目前尚未見開發超級象棋程式對戰象棋大師～期待中。
下圖是象棋著名殘局，紅先勝〈若下緩著則紅敗〉。
象棋殘局01〈網路分享〉
金大俠 於 2016/03/17 10:30回覆
AlphaGo 是 for sale?價格多少？一般棋士會買嗎？
在電腦前下棋，除非三更半夜，沒有生物的氣息，不會很難過嗎？
