2016年1月27日 星期三

Marvin Minsky (1927–2016), Pioneer in Artificial Intelligence

Marvin Minsky (1927–2016), artificial intelligence,The Society of Mind /The Emotion Machine


Marvin Minsky in a lab at M.I.T. in 1968. CreditM.I.T.
Marvin Minsky, who combined a scientist’s thirst for knowledge with a philosopher’s quest for truth as a pioneering explorer of artificial intelligence, work that helped inspire the creation of the personal computer and the Internet, died on Sunday night in Boston. He was 88.
His family said the cause was a cerebral hemorrhage.
Well before the advent of the microprocessor and the supercomputer, Professor Minsky, a revered computer science educator at M.I.T., laid the foundation for the field of artificial intelligence by demonstrating the possibilities of imparting common-sense reasoning to computers.
“Marvin was one of the very few people in computing whose visions and perspectives liberated the computer from being a glorified adding machine to start to realize its destiny as one of the most powerful amplifiers for human endeavors in history,” said Alan Kay, a computer scientist and a friend and colleague of Professor Minsky’s.
Fascinated since his undergraduate days at Harvard by the mysteries of human intelligence and thinking, Professor Minsky saw no difference between the thinking processes of humans and those of machines. Beginning in the early 1950s, he worked on computational ideas to characterize human psychological processes and produced theories on how to endow machines with intelligence.
Professor Minsky, in 1959, co-founded the M.I.T. Artificial Intelligence Project (later the Artificial Intelligence Laboratory) with his colleagueJohn McCarthy, who is credited with coining the term “artificial intelligence.”
Beyond its artificial intelligence charter, however, the lab would have a profound impact on the modern computing industry, helping to impassion a culture of computer and software design. It planted the seed for the idea that digital information should be shared freely, a notion that would shape the so-called open-source software movement, and it was a part of the original ARPAnet, the forerunner to the Internet.
Professor Minsky’s scientific accomplishments spanned a variety of disciplines. He designed and built some of the first visual scanners and mechanical hands with tactile sensors, advances that influenced modern robotics. In 1951 he built the first randomly wired neural network learning machine, which he called Snarc. And in 1956, while at Harvard, he invented and built the first confocal scanning microscope, an optical instrument with superior resolution and image quality still in wide use in the biological sciences.
His own intellect was wide-ranging and his interests were eclectic. While earning a degree in mathematics at Harvard he also studied music, and as an accomplished pianist, he would later delight in sitting down at one and improvising complex baroque fugues.
Marvin Minsky in an undated photo. CreditLouis Fabian Bachrach
Professor Minsky was lavished with many honors, notably, in 1969, the Turing Award, computer science’s highest prize.
He went on to collaborate, in the early ’70s, with Seymour Papert, the renowned educator and computer scientist, on a theory they called “The Society of Mind,” which combined insights from developmental child psychology and artificial intelligence research.
Professor Minsky’s book “The Society of Mind,” a seminal work published in the mid-1980s, proposed “that intelligence is not the product of any singular mechanism but comes from the managed interaction of a diverse variety of resourceful agents,” as he wrote on his website.
Underlying that hypothesis was his and Professor Papert’s belief that there is no real difference between humans and machines. Humans, they maintained, are actually machines of a kind whose brains are made up of many semiautonomous but unintelligent “agents.” And different tasks, they said, “require fundamentally different mechanisms.”
Their theory revolutionized thinking about how the brain works and how people learn.
“Marvin was one of the people who defined what computing and computing research is all about,” Dr. Kay said. “There were four or five supremely talented characters from back then who were early and comprehensive and put their personality and stamp on the field, and Marvin was among them.”
Marvin Lee Minsky was born on Aug. 9, 1927, in New York City. The precocious son of Dr. Henry Minsky, an eye surgeon who was chief of ophthalmology at Mount Sinai Hospital, and Fannie Reiser, a social activist and Zionist.
Fascinated by electronics and science, the young Mr. Minsky attended the Ethical Culture School in Manhattan, a progressive private school from which J. Robert Oppenheimer, who oversaw the creation of the first atomic bomb, had graduated. (Mr. Minsky later attended the affiliated Fieldston School in Riverdale.) He went on to attend the Bronx High School of Science and later Phillips Academy in Andover, Mass.
After a stint in the Navy during World War II, he studied mathematics at Harvard and received a Ph.D. in math from Princeton, where he met John McCarthy, a fellow graduate student.
Intellectually restless throughout his life, Professor Minsky sought to move on from mathematics once he had earned his doctorate. After ruling out genetics as interesting but not profound, and physics as mildly enticing, he chose to focus on intelligence itself.
“The problem of intelligence seemed hopelessly profound,” he told The New Yorker magazine when it profiled him in 1981. “I can’t remember considering anything else worth doing.”
To further those studies he reunited with Professor McCarthy, who had been awarded a fellowship to M.I.T. in 1956. Professor Minsky, who had been at Harvard by then, arrived at M.I.T. in 1958, joining the staff at its Lincoln Laboratory. A year later, he and Professor McCarthy founded M.I.T.’s AI Project, later to be known as the AI Lab. (Professor McCarthy left for Stanford in 1962.)
Professor Minsky’s courses at M.I.T. — he insisted on holding them in the evenings — became a magnet for several generations of graduate students, many of whom went on to become computer science superstars themselves.
Among them were Ray Kurzweil, the inventor and futurist; Gerald Sussman, a prominent A.I. researcher and professor of electrical engineering at M.I.T.; and Patrick Winston, who went on to run the AI Lab after Professor Minsky stepped aside.
Another of his students, Danny Hillis, an inventor and entrepreneur, co-founded Thinking Machines, a supercomputer maker in the early 1990s.
Mr. Hillis said he had so been taken by Professor Minsky’s intellect and charisma that he found a way to insinuate himself into the AI Lab and get a job there. He ended up living in the Minsky family basement in Brookline, Mass.
“Marvin taught me how to think,” Mr. Hillis said in an interview. “He had a style and a playful curiosity that was a huge influence on me. He always challenged you to question the status quo. He loved it when you argued with him.”
Professor Minsky’s prominence extended well beyond M.I.T. While preparing to make the 1968 science-fiction epic “2001: A Space Odyssey,” the director Stanley Kubrick visited him seeking to learn about the state of computer graphics and whether Professor Minsky believed it would be plausible for computers to be able to speak articulately by 2001.
Professor Minsky is survived by his wife, Gloria Rudisch, a physician; two daughters, Margaret and Juliana Minsky; a son, Henry; a sister, Ruth Amster; and four grandchildren.
“In some ways, he treated his children like his students,” Mr. Hillis recalled. “They called him Marvin, and he challenged them and engaged them just as he did with his students.”
In 1989, Professor Minsky joined M.I.T.’s fledgling Media Lab. “He was an icon who attracted the best people,” said Nicholas Negroponte, the Media Lab’s founder and former director.
For Dr. Kay, Professor Minsky’s legacy was his insatiable curiosity. “He used to say, ‘You don’t really understand something if you only understand it one way,’” Dr. Kay said. “He never thought he had anything completely done.”
Correction: January 27, 2016 
An obituary on Tuesday about Marvin Minsky, a pioneer in artificial intelligence, misstated the year he received the Turing Award, computer science’s highest prize. It was 1969, not 1970.

撰文者:愛范兒 發表日期:2016/01/27
美國當地時間 2016 年 1 月 24 日,人工智慧先驅Marvin Minsky(馬文·閔斯基)因腦溢血與世長辭,享年 88 歲。
Marvin Minsky 一生有諸多成就,以下 7 個僅作為引子,希望讓更多人有興趣深入瞭解這位人工智慧研究的奠基人。


1951 年,Marvin Minsky 提出了關於「思維如何萌發並形成」的一些基本理論,並建造了世界上第一個神經元網路模擬器——Snarc(Stochastic Neural Analog Reinforcement Calculator),它能夠在其 40 個「代理」 (Agent)和一個獎勵系統的幫助下穿越迷宮。
在 Snarc 的基礎上,Minsky 還通過綜合利用自己多學科的知識,使機器具備了基於過去行為預測當前行為的能力。
基於 agent 的計算和分散式智慧是當前人工智慧研究中的一個熱點,Snarc 雖然還比較粗糙和不夠靈活,但是人工智慧研究中最早的嘗試之一。

創立 MIT 人工智慧計畫

1956 年,Marvin Minsky 和 John McCarthy 一起發起了被視為人工智慧起點的「達特茅斯會議」,兩人也聯合提出了「人工智慧」的概念。
1959 年,兩者又一同創立了 MIT(麻省理工)人工智慧計畫,這個計畫後來演變成了世界上第一座專攻人工智慧的實驗室——MIT AI 實驗室
有意思的是,除了進行人工智慧研究,MIT AI 實驗室也幫助塑造一種電腦和軟體設計的文化,對於現代計算產業(computing industry)有著深遠影響。它為「電子資訊應該免費獲得」這一理念埋下了種子,這一理念後來協助推展了開源軟體運動。


1969 年,年僅 42 歲的 Marvin Minsky 獲得了電腦科學領域的最高獎項——圖靈獎,他是第一位獲此殊榮的人工智慧學者。
圖靈獎是國際電腦協會(ACM)於 1966 年設立的,又叫 A.M. 圖靈獎,其名稱取自電腦科學的先驅、英國科學家阿蘭 · 圖靈。圖靈一般每年只獎勵一名電腦科學家,有「電腦界的諾貝爾獎」之稱。

出版《The Society of Mind》

1985 年,Marvin Minsky 出版了一本開創性的著作《The Society of Mind》。這部著作提出了「智慧不是任何單獨的機制的產物」這一觀點——Intelligence is not the product of any singular mechanism but comes from the managed interaction of a diverse variety of resourceful agents。
Marvin Minsky 認為,人類實際上就是某種機器,人類的大腦是由許多半自主但不智慧的「代理(agent)」所構成的。他有一句話廣為流傳:「大腦無非是肉做的機器而已(the brain happens to be a meat machine)。」


史丹利·庫柏力克執導《2001太空漫遊》時,專門去請教了 Marvin Minsky,電影裡面的人工智慧電腦 HAL 9000 應該是什麼樣子?
「原來他們有一個裝飾著彩色標籤的電腦。史丹利·庫柏力克問我,您覺得這個怎麼樣? 」在接受《科學發現》雜誌採訪時,Marvin Minsky 說道,「我認為這個電腦實際上應該只是由許多小黑盒子組成,因為電腦需要通過引線來傳遞資訊以知道它裡面在做什麼。」於是庫柏力克把原來的裝飾撤掉,設計了一個簡單的 HAL 9000 電腦。
另外值得一提的是,電影裡有這樣一個場景:HAL 9000 接受 BBC 的訪問,他認為自己「完全不會犯錯」,另一個受採訪的科學家表示 HAL 也會有真實情感。這部電影折射了當時人工智慧專家的一些預測:機器會很快擁有人類水準的智慧。同時,這部電影也引發了對人工智慧或許會變成一件壞事的擔憂。


Marvin Minsky 也是虛擬實境(virtual reality)早期的宣導者。20 世紀 80 年代,Minsky 發表了一篇論文,提出了 Telepresence 遠端控制系統。
Marvin Minsky 設想,人們穿上一個佈滿感測器的、像肌肉一樣的裝置,肩膀、手部、手指的每一個動作都準確無誤地複製到另一個地方的移動機械手柄上。「它允許人體驗某種事件,而不需要真正介入這種事件」。
Minsky 認為,這種遠端作業系統能改變製造、能源和醫院等行業的生產方式。完整的論文點這裡

業界巨星的導師 Marvin Minsky

在 MIT 教出了不少電腦科學的超級巨星,如 Ray Kurzweil(雷·庫茨魏爾),Google 工程總監,同時是未來學家、奇點大學校長;Gerald Sussman,傑出的 AI 研究人員,也是 MIT 的電子工程的教授;Patrick Winston,在 Minsky 教授退休後接管了 AI 實驗室。
雷·庫茨魏爾如此 Marvin Minsky:「在人工智慧、認知心理學、數學、計算語言學、機器人和光學等諸多領域作出了巨大的貢獻,近年來,他一直致力於讓機器具備人類常識推理的能力。對於我來說,他是一位非常值得尊敬的導師。」
附:Marvin Minsky 在 TED 上的演講影片
分享圖來自:Sethwoodworth分享於Wikipedia, cc by 3.0