Lecture 01 introduction to ai

Introduction to Artificial IntelligenceLecture-01

Hema Kashyap

Concerned with the design of intelligence in artificial device.

Term was coined by McCarthy in 1956. Two ideas in the definition

Intelligence Artificial Device

Basic Concept of AI

As Human Ideal Performance

Thought Process /Reasoning Final Manifestation in terms of its Actions

What is Intelligence

• Ability to interact with the real world▫ to perceive, understand, and act▫ e.g., speech recognition and understanding and synthesis▫ e.g., image understanding▫ e.g., ability to take actions, have an effect

• Reasoning and Planning▫ modeling the external world, given input▫ solving new problems, planning, and making decisions▫ ability to deal with unexpected problems, uncertainties

• Learning and Adaptation▫ we are continuously learning and adapting▫ our internal models are always being “updated”

e.g., a baby learning to categorize and recognize animals

What’s involved in Intelligence?

Definition vary along two dimensions

Defining Artificial Intelligence

Thought and Reasoning

Behavior/Actions

Human like Performance

Ideal Performance /Rationality

System that Thinks like

HumanEg: Cognitive

Modeling

System that think RationallyEg: Laws of Thoughts and Logic

System That act like HumanEg: Turing Test

System that act RationallyEg: Rational Agent

It talks about a program that thinks like human

There are two ways to do so: Introspection Through Psychological Experiments

Eg: Program that plays chess like human

Thinks Humanly: The Cognitive modeling approach

A Greek philosopher Aristotle was the first one to codify “Right Thinking” i.e. reasoning process.

His syllogism provided patterns for arguments structures that always yielded correct solutions/conclusions when given correct premises

Eg: Socrates is a man. All men are mortal. Therefore, Socrates is mortal.

This study initiated the field called Logic Two obstacles to this approach:

State problem in the formal terms using logical notations There’s a difference between being able to solve a problem “in

principle” and implementing it in real.

Think Rationally: The “Laws of Thought” process

Proposed by Alan Turing in 1950 to provide a satisfactory definition of AI.

He suggested a test based on in-linguiability from undeniably intelligent entity-human beings.

The computer passes the test if a human interrogator , posing the written questions, cannot tell whether the responses came from a person or not.

But to achieve this a computer would need to possess following capabilities: Natural Language Processing Knowledge representation Automated Reasoning Machine Learning

Acting Humanly: The Turing Test Approach

This test includes the video signal so that the interrogator can test the subject’s perception abilities.

To pass this computer needs Computer Vision: to perceive objects Robotics: to manipulate objects and move

about.

Total Turing Test

Room with operator and huge Chinese literature

Chinese people outside sending in some Chinese texts.

If the operators on looking up the literature able to respond /send text written in front of the text received, then person outside believes operator knows Chinese.

Its just the matter of “Translating”.

This doesn’t mean that the person understands semantics of the language.

So, Cognition and Understanding is different thing

Chinese Room Test

Agent(that acts), and a computer program agent is more than just a mere program, the one That operates under autonomous control Perceive their environment Persisting over a prolonged time period Adapting for change Being capable of taking an another’s goal

A Rational Agent is the one that acts so as to achieve the best outcome or when there is uncertainty, the best expected outcome.

Act Rationally Reason Logically Draws Conclusions Acts on that conclusion

Acting Rationally: The Rational agent Approach

Intelligent Agents need to be able to do both “MUNDANE” and “EXPERT” task.

Mundane Task: planning route actively, recognizing, communicating etc.

Expert Task(needs domain specific knowledge): medical diagnosis, mathematical problem solving etc.

Defining: Typical AI problem