《人工智能-一种现代的方法》阅读笔记
Chapter 1. Introduction
- Definition of AI
In which we try to explain why we consider intelligence to be a subject most worthy of study, and in which we try to decide what exactly it is, this being a good thing to decide before embarking(着手).
| Thinking Humanly ② | Acting Humanly ① |
| Thinking Rationally ③ | Acting Rationally ④ |
① To acting humanly to pass the Turing test include plenty capabilities to work on:
Natural language processing
Knowledge representation
Automated reasoning
Machine learning
Computer vision
Robotics
——Six disciplines
② To thinking humanly, AI and cognitive science continue to fertilize each other, most notably in computer vision, which incorporates neurophysiological evidence into computational models.
③ To thinking rationally, Greek philosopher’s study of laws of thought initiated the field called logic.
④ To acting rationally, it takes a rational agent who is the one that acts so as to achieve the best outcome or, when there is uncertainty, the best expected outcome.
2. Foundations of AI
① Philosophy: there are empiricism, rationalism, dualism and materialism.
② Mathematics: decidability, computability and tractability are important to understand computation. The theory of NP-completeness and Bayes’ rule play crucial roles in modern AI world.
③ Economics: Decision theory combines probability theory with utility theory. The topic of how to make rational decisions when payoffs from actions are not immediate but instead result from several actions taken in sequence is pursued in the field of operations research. The idea of making decisions that are “good enough” rather than calculating an optimal decision gave a better description of actual human behavior, which is called “satisficing”.
④ Neuroscience: Even with a computer of virtually unlimited capacity, we still would not know how to achieve the brain’s level of intelligence. EEG and fMRI gave unprecedently detailed images of brain activity, enabling measurements that advances in single-cell recording of neuron activity.
⑤ Psychology:
⑥ Computer engineering
⑦ Control theory and cybernetics: Modern control theory, especially the branch known as stochastic optimal control, has as its goal the design of systems that maximize an objective function over time. This roughly matches our view of AI: designing systems that behave optimally.
⑧ Linguistics:
3. The History of Artificial Intelligence
The gestation of artificial intelligence was during the 1940s and 1950s, and was born in accordance with a workshop host at Dartmouth in 1956.
4. The State of The Art
Robotic vehicles, speech recognition, autonomous planning and scheduling, game playing, spam fighting, logistic planning, robotics and machine translation.
5. Summary
Chapter 2 Intelligent Agents
1. Agents and Environments
Sensors of agents percept the environment and its actuators give series of actions to update the environments’ states.
2. Good Behavior: The Concept of Rationality
A rational agent is expected to maximize its performance measure. It improves its autonomy by learning, though don’t have omniscience.
3. The Nature of Environments
The hardest case is partially observable, multiagent, stochastic, sequential, dynamic, continuous, and unknown environment.
4.The Structure of Agents
There are simple reflex agents, Model-based reflex agents, goal-based agents and utility-based agetns.