Introduction to Artificial Intelligence
What could the term "Artificial Intelligence" possibly mean, if anything? Is it worth studying?
Definitions
There are hundreds of definitions of artificial intelligence. Most contain a bias as to whether the writer of the definition sees AI as dealing with thinking versus acting, and whether they see it as trying to model humans or capturing intelligence (rationality) abstractly.
| Humanly
| Rationally
|
Thinking
| Thinking humanly — cognitive modeling. Systems
should solve problems the same way humans do.
| Thinking rationally — the use of logic. Need to
worry about modeling uncertainty and dealing with
complexity.
|
Acting
| Acting humanly — the Turing Test approach.
| Acting rationally — the study of rational agents: agents
that maximize the expected value of their performance
measure given what they currently know.
|
Some definitions
- The study of agents that receive percepts from the environment and perform actions. (Russell and Norvig)
- The science and engineering of making intelligent machines, especially intelligent computer programs (John McCarthy)
- The ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings (Encyclopædia Britannica)
- The study of ideas to bring into being machines that respond to stimulation consistent with traditional responses from humans, given the human capacity for contemplation, judgment and intention (Latanya Sweeney)
- The scientific understanding of the mechanisms underlying thought and intelligent behavior and their embodiment in machines (American Association for Artificial Intelligence)
- A branch of science which deals with helping machines find solutions to complex problems in a more human-like fashion (AI depot)
- A field of computer science that seeks to understand and implement computer-based technology that can simulate characteristics of human intelligence and human sensory capabilities (Raoul Smith)
Eight more can be found here. And a handful more here.
Concerns
AI has both scientific (nature of intelligence) and engineering (design
and production of intelligent agents) aspects.
AI Science: The Nature of Intelligence
AI researchers study the nature of intelligence. They
don’t (necessarily) try to build androids because
- It was the study of the principles of aerodynamics,
not the attempt to make mechanical birds, that enabled human flight.
- We already know how to make new humans anyway (sorry, no
hyperlinks here).
Intelligence involves sensing, thinking, and acting.
SENSING
| THINKING
| ACTING
|
Translation of sensory inputs (percepts)
into a conceptual representation
- Computer Vision
- Speech Recognition
- Language Understanding
| Manipulation of the conceptual
representation
- Knowledge Representation
- Problem Solving/Planning
- Learning (making improvements based on the results of past actions)
| Translation of intent into (physical) actions (reflexive
or deliberative)
- Robotics
- Speech and Language Synthesis
|
AI Engineering: Intelligent Agents
An agent is something that senses and acts.
- Agents can be organic, robotic, or pure software.
- Not all agents are intelligent.
- Intelligent agents are agents that
can operate autonomously in complex environments.
- AI is concerned with the production of intelligent agents.
Areas of Study
In no particular order, nor with any thought of completeness, here are
a few:
- Vision
- Speech recognition
- Robotics
- Problem Solving
- Searching a Solution Space
- Planning
- Learning
- Natural Language Processing
- Natural Language Understanding
- Knowledge Representation
- Automated Reasoning
- Inference, both in monotonic and non-monotonic logic
- Common Sense Reasoning
- Uncertainty and Probability
- Genetic Programming
- Artificial Life
- Ontology
- Epistemology
- Expert Systems
- Solving problems with no tractable deterministic algorithmic
solutions
Influences
From Section 1.2 of Russell and Norvig:
- Philosophy considers the nature of knowledge, thought, and learning
- Mathematics considers the notions of formal logic, algorithms,
computational complexity, and probability
- Economics studies how agents attempt to maximize their own
well-being, even when given uncertain information and in the presence
of allies and adversaries
- Neuroscience studies the workings of the human brain
- Psychology studies how humans and animals think and act
(process information)
- Linguistics deals with language in a formal-enough
way that it can be processed by machine
- Computer Engineering looks to increase the efficiency
of computing devices
- Control Theory and Cybernetics consider how autonomous
machines can operate
Brief History
Highlights:
- 1940’s — Interest in neurons, neural networks and their
relationship to mathematics and learning
- 1950 — Turing’s paper
- 1956 — Dartmouth conference
- 1950’s and 1960’s — enthusiasm and optimism; big promises
- Late 1960’s and 1970’s — Realization that further progress
was really hard; disillusionment
- 1980’s — Expert Systems, neural networks, etc.; AI now a
little different; quiet successes
- 1990’s to present — Intelligent agents
- 2000’s — robot pets, self-driving cars
More history:
Applications
Today AI is in:
- Game playing (Chess, Go, Risk, Bridge, Checkers, ...)
- Systems that read handwritten addresses to speed mail sorting
- Search Engines
- Theorem Proving
- Cars (stability traction, braking assist, driving, ...)
- Aircraft autolanders
- Medical Diagnosis
- Expert Systems
- Information Retrieval Systems
- Story writers, poetry writers, ...
- Music Composition
- Annoying auto-correct agents in word processors
- Crisis management
- Space Exploration
- Finance
- Retailing
- Manufacturing
- Inventory Control
- Pharmaceutical Research
- Genetic Research
- (Micro)Surgery
- Insurance Underwriting
- Environmental Monitoring
- Protein Structure Determination
- Scheduling Systems
- Assisted Living Support
- Dispensing Legal Advice
- Essay Evaluation
- Detection of Steganography
- Cryptanalysis
- Translation
- Military Planning
- Surveillance
- Traffic Control
See Also