History ​
1956
First Use of the termArtificial Intelligence
at the dartmouth summer research project.1956-1974
: "golden years", shakey robot.
AI Winters and Spring ​
AI winters happen when inflated expectations of AI results are created then there is a failure to meet those expectations. This results in reduced funding for the field.
Problems:
- modelling commonsense knowledge.
- intractability.
- frame problem. (model what does not change in the environment)
The first AI winter happened from 1974-1980
Then from 1980-1987
: expert systems + rule based systems for modelling the knowledge of experts. + brittle and difficult to maintain. Then came the 2nd AI winter from 1987-1993
Then from 1993-2011
+ Increase in computing power. 💻 + Intelligent agents 🤖 + Decision Theory Finally from 2011-Now
+ Big data enables Machine Learning.
AI Pendulum ​
- Oscillation between high expectations/funding and the opposite.
- Oscillation of techniques: symbolic vs sub-symbolic/data oriented
Currently attention is being brought to bridging AI subfields to obtain better functioning AI.