In the sixth order of the series, we look at Marvin Minsky, one of the pioneers who laid the foundation for artificial intelligence. If you haven't checked out the past content of the series yet, I recommend reading it first.
“Humans are thinking machines.”
Marvin Lee Minsky (1927—2016), one of the major pioneers who laid the foundation for early artificial intelligence, said this. I would like to introduce his story, who left important achievements through his lifelong research activities with the belief that he can actually make machines that think like humans.
Minsky was born in 1927 in New York and attended the Bronx School of Science. Franklin Rosenblat*, who later confronted him with Perceptron (the fourth main character in this series), was his one-year junior. Since then, he has received bachelor's and doctorate degrees in mathematics from Harvard and Princeton, and has continued his active research activities by working as a professor at MIT.
He was a pioneer of artificial intelligence and an educator who was active in various fields such as mathematics, cognitive science, computer science, and robotics while cultivating outstanding students. Of course, this article will focus on achievements related to artificial intelligence. He was a versatile person known for even playing the piano and composing songs. **
SNARC, the first neural network computer
In 1951, he succeeded in developing the first neural network computer (SNARC, StochasticNeural Analog Calculator) with Dean Edmonds, who was a doctoral researcher at Princeton University. Artificial intelligence had already begun in the vacuum tube era a long time ago. However, in a very different form from today's computers, it was a network system with about 3,000 vacuum tubes and about 40 Hebb (Hebb) synapses randomly connected. ***
SNARC was a neural network machine that could learn through trial and error. In other words, it was the result of developing his concerns about thinking machines. Minsky's mentor at the time was skeptical about this machine, but it is said that von Neumann praised it, saying that it is very advanced and will be very useful one day. ****
This machine was clearly one of the pioneering experiments in the field of artificial intelligence. (Ironically, Minsky later strongly criticized the problem of neural networks, which led to a harsh period called the winter of artificial intelligence.)
Dartmouth conference announcing the birth of artificial intelligence
Minsky and John McCarthy**** (the third protagonist in this series) were good friends. With these two at the center, scholars such as Nathaniel Rochester and Claude Shannon took center stage and promoted the Dartmouth Conference. In the summer of 1956, pioneers of artificial intelligence, who were still young at the time, gathered for about two months.
The purpose of the meeting was to discuss the construction of machines considered intelligent. However, compared to its meaning and importance, it is said that there were not many actual results at the time. However, through this meeting, “Artificial Intelligence (AI)” made its name known in history for the first time.
This is how artificial intelligence was born in Dartmouth in the summer of 1956.
head-mounted display
Minsky and McCarthy started an artificial intelligence project by establishing a laboratory that developed from MIT into the current Computer Science and Artificial Intelligence Laboratory (CSAIL). In particular, in the early days, they developed robotic arms that sense touch or lift objects, and visual scanners that act as robot eyes. In other words, I studied how machines can have human senses, cognitive abilities, and intelligence. ****
The “Head MountedDisplay (HMD)” used in virtual reality and games today was also created at this time. A small display such as a liquid crystal was installed close to both eyes to project stereoscopic images using parallax. The direction facing the user's head is detected with a gyro sensor, etc., and images corresponding to movement are emphasized, making it possible to experience as if they were in a three-dimensional space. ****
Perceptron's limitations, XOR problems
Perceptron, an early model of today's artificial neural networks and deep learning, was invented by Frank Rosenblatt in 1957. For the first time, the principle that the neural network of the human brain works was implemented with a mechanical device. Furthermore, the demonstration of image recognition at the time was successful, and great social interest and research support were focused on neural network research.
However, the limitations of Perceptron, which attracted so much attention, were revealed by Minsky. Unlike Rosenblatt, who chose to study neural networks, Minsky was studying expert systems. This was a confrontation between two geniuses, but it was also a confrontation between neural network-based connectionism and rule-based symbolism, which are two major trends in artificial intelligence.
<퍼셉트론즈>Marvin Minsky and his colleague Seymour Pepper publish Proving the Limits of Perceptron. Perceptron points out that XOR computation is not possible. However, it was later revealed that this is a problem that can be solved through multi-layer perceptrons and algorithm improvements. *****
There is no such thing in history, but it is a pity that if these two had cooperated with each other, a faster development of artificial intelligence would have been possible.
The framework of intellectual activity
In 1974, he published “A Framework for Intellectual Knowledge (A Framework for Intellectual Knowledge),” one of his important achievements. It was a theory that incorporated psychological methods into artificial intelligence research. This was an innovative idea that expressed human knowledge in a data structure called a frame and saw intellectual activities such as language understanding, pattern recognition, and problem solving as an interaction between data input from outside and an internal frame. ******
Mind society
In 1985, he published his flagship book, The Society of Mind (The Society of Mind). This book was his answer to the age-old question “How does the human mind work?” It contains his outstanding insights covering not only artificial intelligence, but also the humanities and natural sciences such as cognitive science, psychology, and philosophy.
Through this book, Minsky says, “Intelligence is not a product of a single mechanism, but rather comes from the managed interaction of 'acting elements' with various abilities.” In other words, our brain is a combination of imperfect parts that come together to perform complex functions like a single society. This theory gave us insight and inspiration about how the brain works and how humans learn.
“What kind of magic makes us intelligent? The trick is that there are no tricks. The power of intelligence comes not from any single perfect principle, but from our vast diversity.” (Marvin Minsky, Society of Mind p.308) *******
While finishing
Minsky developed artificial intelligence with the belief that “humans are machines that think.” Our brain is a form where many nerve cells are connected, and not each cell has intelligence. In the end, I believed that it was possible to create a machine that thinks like a human, or artificial intelligence, depending on how it is connected.
His lifelong challenges and achievements have had a tremendous impact on the development of artificial intelligence. He was a decisive figure in making it possible for computers that could simply stay in computational machines to evolve into artificial intelligence. He passed away from this world in 2016, but the era where machines that think like the humans he dreamed play an active role is getting closer and closer to reality.
References
[1] http://wiki.hash.kr/index.php/마빈_민스키
[2] https://ko.wikipedia.org/wiki/SNARC
[3] http://wiki.hash.kr/index.php/마빈_민스키
[4] [AI character exhibition] Marvin LeeMinsky (Marvin LeeMinsky) https://www.newsmc.net/news/articleView.html?idxno=683
[5] 'Pioneer of Artificial Intelligence' Marvin Minsky, who is he https://www.dongascience.com/news.php?idx=10001
[6] http://www.aistudy.co.kr/pioneer/Minsky.M.htm
[7] Robot, can you write an obituary of your Creator? http://h21.hani.co.kr/arti/world/world_general/41163.html
[8] [Footsteps] 'Father of Artificial Intelligence' Marvin Minsky https://biz.chosun.com/site/data/html_dir/2016/01/27/2016012700791.html
[9] Marvin Minsky and the Artificial World https://m.blog.naver.com/cadline/18075645
[10] [New Edition] The Society of Mind (The Society of Mind) http://www.futurekorea.co.kr/news/articleView.html?idxno=115058
[11] Involved is a Big Suitcase, A Talk with Marvin Minsky https://www.edge.org/3rd_culture/minsky/index.html
[12] (Interview Video) Interview with Marvin Minsky, 1990 https://youtu.be/DrmnH0xkzQ8
[13] (MIT lecture video) Introduction to 'The Society of Mind' https://youtu.be/-pb3z2w9gDg
Good content to watch together
[AI Story] Key Figures of Artificial Intelligence (1) Alan Turing[AI Story] The decisive figures of artificial intelligence (2) Walter Fitz, the founder of deep learning[AI Story] The decisive figures of artificial intelligence (3) John McCarthy, the founder of artificial intelligence[AI Story] Key Figures of Artificial Intelligence (4) Rosenblatt, pioneer of deep learning in artificial intelligence[AI Story] The decisive figures of artificial intelligence (5) A fantastic combination, Simon and Newel