HackMD
人工智慧導論
目錄
Chapter 1：
認識AI
- Systems that think like humans
- Systems that think rationally
- Systems that act like humans
- Systems that act rationally
屬於光譜，沒有標準答案
Turing Test
Suggested major components of AI
- knowledge
- reasoning
- language
- understanding
- learning
Problem of Turing Test
- not reproducible
- not constructive
- not amenable to mathematical analysis
不可重現、沒有建設性、不適合數學分析
The Computer would need the following capabilities
- Natural language processing (NLP)
- Knowledge representation
- Automated reasoning
- Machine learning (ML)
To pass the total Turning Test, a robot will need
- Computer vision
- robotics
Thinking humanly: Cognitive Science
Thinking rationally: Laws of Thought
Problems：
- Not all intelligentbehavior is mediated by logical deliberation
- What is the purpose of thinking? What thoughts should I have out of all the thoughts (logical or otherwise) that I could have?
Right Thinking
- Socrates is a man
- All man are mortal
- Socrates is mortal
Acting Rationally
Rational behavior: doing the right thing
AI Prehistory
philosophy
- logic
- methods of reasoning
- mind as physical system
- foundations of learning
- language
- rationality
Thinking
- Can formal rules be used to draw valid conclustions?
- How does the mind arise from a physical brain?
- Where does knowledge come from?
- How does knowledge lead to action?
Mathematics
- formal representation and proof
- algorithms
- computation
- (un)decidability
- (in)tractability
- probability
Psychology
- adaptation
- phenomena of perception and motor control
- experimental techniques (psychophysics, etc.)
Thinking
- How do humans and animals think and act?
Economics
- formal theory of rational decisions
Thinking
- How should we make decisions in accordance with our preferences?
- How should we do this when others may not go along?
- How should we do this when the payoff may be far in the future?
Linguistics
- knowledge representation
- grammar
Thinking
- How does language relate to thought?
Neuroscience Control theory
- plastic physical substrate for mental activity
Thinking
- How do brains process information?
Control theory
- homeostaic systems
- stability
- simple optimal agent designs
History
The inception of AI (1943 - 1956)
- McCulloch & Pitts
- Hebbian learning
A does of reality (1966 - 1973)
- Main reasons for this failure
- 缺乏對事物的認知
- 當時AI系統主要基於知情內省，而沒有可靠的解決方案(如演算法)
- Machine evolution(now call genetic programming)
Expert systems (1969 - 1986)
- weak methods
- certainty factors
- In 1981, the Japanese government announced the "Fifth Generation" project
- AI Winter
many companies fell by the wayside as they failed to deliver on extravagant promises
The return of neural networks (1986 - present)
- Back-pro
Probabilistic reasoning and machine learning (1987 - present)
Big data (2001 - present)
This has led to the development of learning algorithms specially designed to take advantage of very large data sets.
Deep learning (2011 - present)
- 2012 ImageNet Competition
- Deep learning relies heavily on powerful hardware
Risks and Benefits of AI
"First solve AI, then use AI to solve everything else."
Demis Hassabis, CEO of Google DeepMind
- Benefits
- Decrease repetitive work
- Increase production of goods and services
- Accelerate scientific reasearch
disease cures, climate change and resource shortages solutions
- Risks
- Lethal autonomous weapons
- Surveillance and persuasion
- Biased decision making
- Impact on employment
- Safety-critical applications
- Ccyversecurity threats
Chapter 2：Intelligent Agents
Outline
- Agents and environments
- Rationality
- PEAS (Performance measure, Environment, Actuators, Sensors)
- Environment types
- Agent types
Agents and Environments
Agents include humans, robots, softbot, thermostats, etc.
An agent can be anything that can be viewed as perceiving its environment through sensors and acting upon that environment through actuators.
The agent function maps from percept histories to actions:
f:P* -> A
The agent program runs on the physical architecture to produce f
Agents interact with environments through sensors and actuators
Rationlity
- The performance measure that defines the criterion of success
- The agent's prior knowledge of the environment
- The actions that the agent can perform
- The agent's percept sequence to date
PEAS
To design a rational agent, we must specify the task enviironment Consider, e.g., the task of designing an automated taxi:
- Performance measure：safety, destination, profits, legality, comfort…
- Environment：streets, traffic, pedestrians, weather…
- Actuators：steering, accelerator, brake, horn, speaker/display…
- Sensors：video, accelerometers, gauges, engine sensors, keyboard, GPS…
Chapter 3：Solving Problems by Searching
Outline
- Problem-solving agents
- Example Problems
- Problem formulation
- Search Algorithms
- Uninformed Search Strategies
- Informed (Heuristic) Search Strategies
- Heuristic Functions
Problem-solving agents
- seq
an action sequence, initally empty
- state
some description of the current world state
- goal
a goal, initally null
- problem
a problem formulation
