In the second order of the series, we will learn about Walter Fitz, who opened up the origin of deep learning through research that models the human brain. If you haven't watched past content in the series yet, I recommend checking it out first.
[AI Story] Key Figures of Artificial Intelligence (1) Alan Turing
Walter HarryPitts, Jr. (1923.04.23 to 1969.05.14) made an important achievement of mathematically modeling the working principles of human neurons for the first time. It can be said to be a pioneer who pioneered artificial neural networks, machine learning, and even deep learning research. Now let's take a look at his life and research, which played an important role in the early history of artificial intelligence.
A genius boy in a poor street
<굿 윌 헌팅>A film reminiscent of Walter Fitz's life, sources
Walter Fitz was born into a poor family in Detroit, USA. Furthermore, he was a dwarf boy who was abused by his father and bullied by local children. His life reminds me of the story of an unlucky genius boy in the movie “Good Will Hunting (Good Will Hunting)” directed by Gus Van Sant (Gus Van Sant).
In 1935, 12-year-old Fitz found a book destined to change his life at a library that had escaped local bullies. It was “Principia Mathematica (Principia Mathematica)” written by Bertrand Russell and Alfred Whitehead. He read all of this difficult book, which spans hundreds of pages, over several days.
Furthermore, to my surprise, I found some errors in the book's content. Then I sent a letter to Russell, the author of the book, about it. Impressed by this, Russell tried to bring this genius boy to the UK to teach, but he was unable to enroll him at Cambridge Graduate School when he was only 12 years old.
Warren McCulloch and Walter Fitz, sources
However, as a result of this relationship, Fitz learned logic and mathematics, and had a valuable opportunity to meet many people. Neuroscientist Warren Sturgis McCulloch (Warren Sturgis McCulloch), who later collaborated with him, was one of them. McCulloch acted as a guardian so that Fitz, who was wandering around while doing cleaning and other errands, could stay with him.
The beginning of deep learning, the McCulloch-Fitz model
The McCullock-Fitz model shown with thesis and <신경 활동에 있어서 관념의 논리적 미적분> propositional logic, sources
Afterwards, the two will present the results of their joint research. This is the paper “A Logical Calculus of Ideas Impaired in Nervous Activity (A Logical Calculus of Ideas Impaired in Nervous Activity) in Neural Activity. Through this paper, the two explain the function of neurons with a binary logical model consisting of 0 and 1.
The following is the summary section of this paper.
“Because of the 'all-or-nothing (all-or-nothing) 'nature of neural activity, the work of the nervous system and the relationships between them are treated as propositional logic (propositional logic). The behavior of any network can be described from this perspective.... short note... For a logical expression that aims to satisfy certain conditions, we can find a network that acts in the way it describes. ” *
This was the first study to mechanically model the human brain. The function of each neuron is very simple, but it was this paper that laid the foundation for a neural system that performs complex calculations by connecting them to each other. This later became the foundation that made modern computers with binary output possible.**
The impact and importance of this paper on future generations is huge. Fifteen years after the McCullock-Fitz model was announced, it led to the concept of a “perceptron.” This perceptron evolved into a neural network, and the neural network evolved into deep learning. **
An unlucky life and great accomplishments
Fitz began his research at MIT and attracted attention as a promising young scientist. Unfortunately, his life didn't go smoothly after that. I ended my short, lonely life after suffering from depression and alcoholism alone.
Fitz had a disadvantaged childhood. I think the time I spent with McCulloch was probably a very happy time. Although he lived such an unlucky life, his outstanding achievements have led to the development of deep learning today.
* Quote http://www.aistudy.co.kr/neural/McCulloch_Pitts.htm
** Citations and references http://wiki.hash.kr/index.php/월터_피츠#.EB.94.A5.EB.9F.AC.EB.8B.9D.EC.9D.98_.EC.8B.9C.EC.B4.88
References
[1] https://en.wikipedia.org/wiki/Walter_Pitts
[2] http://afflictor.com/2015/02/05/they-would-create-the-first-mechanistic-theory-of-the-mind/
[3] http://wiki.hash.kr/index.php/월터_피츠
[4] https://ko.hrvwiki.net/wiki/Walter_Pitts
[5] http://www.aistudy.co.kr/neural/McCulloch_Pitts.htm
[6] The Man Who Missed to Redeem the World with Logic https://nautil.us/issue/21/information/the-man-who-tried-to-redeem-the-world-with-logic
[7] 70th Anniversary of Publication: WarrenMcCulloch & Walter Pitts - A Logical Calculus of the Ideas Impossible InNervous Activity https://www.semanticscholar.org/paper/70th-Anniversary-of-Publication%3A-Warren-McCulloch-%26-Posp%C3%ADchal-Kvasnicka/50044ca6ad13e2e852de81237c4939e5973b8982
Good content to watch together
[AI Story] Crucial Moments of Artificial Intelligence 1[AI Story] Key Figures of Artificial Intelligence (1) Alan Turing
