This article is the learning excerpt and notes recorded by the blogger for personal study, research or appreciation when he is studying in artificial intelligence and other fields, and based on the blogger’s understanding of artificial intelligence and other fields. I will correct it immediately after pointing it out, and I hope for your understanding. Articles are categorized in the study excerpts and notes column :
Study Excerpts and Notes (3) --- " Some Thoughts in Artificial Intelligence Research "
Thoughts on Artificial Intelligence Research
Source of original text/paper:
Topic: " Thoughts on Artificial Intelligence Research "
Author: Zhao Lu
Time: 2019-04-04
Source: Human-Computer and Cognition Lab
The opening of the Dartmouth Conference in 1956 announced the birth of the discipline of artificial intelligence. In general, the overall progress of artificial intelligence can be divided into four stages: the incubation stage, the initial development stage, the accumulation stage and the vigorous development stage.
At present, there are several existing bottlenecks in the field of artificial intelligence:
That is, machines are still unable to form common sense, motivation and intelligent decision-making
The direction of development of artificial intelligence:
(1) Swarm Intelligence
When many group living creatures forage for food and escape, the individuals often do not have complex behaviors, but the group behaviors show complex intelligent phenomena.
The five principles of swarm intelligence are proximity principle, quality principle, diversity response principle, stability principle and adaptability principle.
The more famous swarm intelligence includes ant colony optimization , particle swarm optimization algorithm and artificial fish swarm algorithm .
(2) Soft computing
Proposed by Zadeh in 1994, it aims to solve the inaccurate and uncertain direction and content of traditional artificial intelligence computing methods.
Computing is divided into two categories: hard computing and soft computing . The concept of soft computing is widely accepted.
The main content includes artificial neural network , genetic algorithm , fuzzy logic and other theories and methods.
(3) Hybrid intelligent system
As practical applications become more and more complex and data dimensions become higher and higher, traditional machine learning has become difficult to solve existing problems.
If two or more methods are combined , it can achieve the effect of maximizing strengths and avoiding weaknesses.
Table 1: Comparison of advantages and disadvantages of several common methods
| expert system |
fuzzy system |
Neural Networks |
genetic algorithm |
|
| knowledge representation |
very good |
good |
not good |
very bad |
| Tolerance of Uncertainty |
very good |
good |
good |
good |
| Tolerance of imprecision |
not good |
good |
good |
good |
| adaptability |
not good |
very bad |
good |
good |
| learning ability |
not good |
not good |
good |
good |
| interpretive ability |
good |
good |
not good |
very bad |
| Knowledge Discovery and Data Mining |
not good |
very bad |
good |
very good |
| maintainability |
not good |
very good |
good |
very good |
article comments
Focus on the difference between human-machine decision-making mechanism , starting from the underlying algorithm, integrate human motivation, expectation mechanism and decision-making mechanism, and gradually form a decision-making mechanism with independent expectations and motivation .
If the article is inappropriate and incorrect, please understand and point it out. Since some texts and pictures come from the Internet, the real source cannot be verified. If there is any dispute, please contact the blogger to delete it. If you have any errors, questions or infringements, please leave a comment to contact the author, or follow the VX public account: Rain21321, and contact the author.