The counterattack of "neural networks": an illustration of the 80-year history of AI struggle
Release time:
2018-11-29 09:26
Source:
The two main factions are:
In the eyes of symbolists, artificial intelligence should imitate human logic to acquire knowledge, while connectionists pursue big data and training to learn knowledge.
Factional struggles and two AI crises
As early as before the Dartmouth conference, Turing proposed the "Turing machine" such a cutting-edge concept of artificial intelligence. During the early decades of the struggle, the connectionist school consistently led its opponents in citations.
Don't look at the glory of machine learning that pursues "connectionism". In the early years, they were despised by another faction-"symbolism" for a long time.
In the early 1960 s, the U.S. Defense Advanced Research Projects Agency (DARPA) invested millions of dollars in the field of AI, and artificial intelligence also ushered in the first golden period of development.
The First AI Crisis
That changed in 1969, when Marvin Minsky, the representative of "symbolism", wrote a book called "Perceptron" (Perceptron), which ended up killing neural networks.
Related News
2021-07-07