Welcome to Edukum.com


Problem solving

Search Techniques

Knowledge representation, Inference and Reasoning

Structured Knowledge Represent

Machine Learning

Applications of AI

Definition of Artificial Intelligence

The meaning of the intelligence is the capacity to acquire and apply the knowledge. The intelligence is regarded as the faculty of thought and reason. It is the ability to comprehend in order to understand and make a profit from experience. Intelligence can also be a secret information especially about an actual or potential enemy.An agency, staff, or office employed may be involved in gathering such information and use them accordingly.
Based upon the above details, we can say that artificial intelligence is all about the science and engineering necessary to create artifacts that can acquire knowledge which means it can learn and extract knowledge and reason with that knowledge (leading to doing tasks such as planning, explaining, diagnosing, acting rationally, etc.).

Formal Definitions

Barr and Feigenbaum
“Artificial Intelligence is the part of computer science that is concerned with designing intelligent computer systems, that is the systems that exhibit those characteristics which we associate with intelligence in human behavior.”

Elaine Rich
“AI is the study of how to make computers do things at which, at the moment, people are better”.
Artificial Intelligence (AI) can simply be understood as the human-like intelligence exhibited by a machine. The field of AI is inter-disciplinary in which a number of sciences and professions converse such as computer science, psychology, linguistic, neuroscience etc. Artificial intelligence can be defined in terms of following categories:

  1. Acting like Human
    The art of creating machines that perform functions which require intelligence when performed by people is understood as acting like a human. This approach is also called as turing test approach which is not interested in how you got the result rather it is concerned with the similarities to what human behavior are. The turing test measures the performance of intelligence machine against that of human being, Factors required to pass the turing test are:
    Natural language processing (To make communication easier).
    Knowledge representation (To store information).
    Automated reasoning (To use stored information and draw a new conclusion).
    Machine learning (To adopt with new circumstances).

    For example: There is a real Chinese guy in a room who is trying to communicate with an unknown agent in the next room in his own language even though the person in the next person may not necessarily be Chinese. Therefore, the Chinese think that someone in the next room is also Chinese.

  2. Thinking like Human
    This refers to the automation of activities that we associate with human thinking activities such as decision making, problem-solving, learning etc. This phenomenon is also known as the cognitive based approach in which the focus is not just on the behavior and the input but the output also looks at reasoning process. This approach requires understanding how human thinks. Different ways to determine how human thinks are inspections, experience, current perception etc. Example: General problem solver.

  3. Thinking like Rational Agent
    The study of computations that make it possible to perceive, reason and act is meant by thinking like the rational agent. This is also known as Laws of Thought Approach. A system is said to be rational if it does the right thing for what it knows. It attempts to codify the normative premises in terms of logic, driving the result as a right thinking. Logic provides precise notations or statements about all kinds of things and relations between them. For example: For the premises, All men are the student. Ram is a man. The result is: Ram is a student.

  4. Acting Rationally
     Computational Intelligence is the study of the design of intelligence agent. This model is also known as rational agent approach. The rational agent acts to achieve the best outcome when there is the uncertainty of the best expected outcome. A rational agent is more general than laws of thought approach which emphasize on correct inference where correct inference is useful but not necessary for achieving rationally. For example: reflex agent.

An agent perceives data from the environment with the help of a sensor. Mapping is done by the agent by writing a certain program. Finally, the actuators act upon the environment.


Types of Agents

  1. Simple Reflex Agent
    These are simplest type of agent which selects action on the basis of current perceptions. For example: In the human being, blinking of an eye when anything approaches.

  2. Model Based Agent
    The effective way to handle the environment function is to keep track of the facts of the world. Agent should maintain the internal states depending upon the perceived history and reflect the result which is not only based on current state. For example: switch ON the street light in the evening and off in the morning.

  3. Goal Based Agent
    Knowing about the current state of the environment is not always enough to decide what to do next so we need to store more goal information. Although the goal-based agent is less efficient it is more flexible because the knowledge that is provided is decided explicitly and goal can be changed earlier.

  4. Utility-Based Agent
    Utility-based agent refers to choosing the path to obtaining the goal that is best among probable path. It associates the utility function with each state. The utility function is a function that maps a state or sequence of the state into a real number that represents the degree of happiness of agent.

As we discussed the agents we need to know that those agents have to face different situations. The environment can simply be understood as the situation that an agent faces. Types of environment are given below:

  1. Deterministic vs Stochastic
    The deterministic environment is said to be an environment where the next state is fully determined by the current state and the action executed by the agent else the environment is called as a stochastic environment. For example: Playing chess is deterministic while medical analysis is stochastic.

  2. Static vs Dynamic
    The static environment does not change its state while the dynamic environment does. Static environment is easy to deal with than the dynamic environment. In dynamic environment agent always probe about what to do in a particular situation. For example: Crossword puzzle is static while driving a taxi is dynamic.

  3. Fully observable vs Partially observable
    When the state of the environment is accessed completely at any time then it is called fully observable environment otherwise it is the partially observable environment. In the fully observable environment, the sensor will detect all aspects of actions which is relevant to the performance measure. For example: Image analysis is fully observable while the medical diagnosis is partially observable.

  4. Discrete vs Continuous
    If there are limited number of distinct and clearly defined perceptions and actions then such environment is called discrete environment otherwise it is called a continuous environment. For example: Playing chess is discrete because there are fixed numbers of possible moves in each turn whereas automated taxi-driver is continuous.

  5. Episodic vs Sequential
    At each episode an agent performs a single action and next action does not depend on upon the previous action. Such type of environment is called episodic. Sequential environment depends on upon actions and previous actions. The episodic environment is simpler than the sequential environment because in the sequential environment the agent has to maintain the state of the environment of previous episodes. For example: Playing chess or crossword puzzle is sequential and vacuum cleaning is episodic.

  6. Single-agent vs Multi-agent
    If an environment involves only one agent then it is said to be a single agent. In multi-agent environment two or more than two agents are present. For Example: Crossword puzzle involves only one agent. Chess-playing involves two agents.


  1. Elaine Rich, Kevin Knight 1991, "Artificial Intelligence".
  2. Nilsson, Nils J. Principles of Artificial Intelligence, Narosa Publishing House New Delhi, 1998.
  3. Norvis, Peter & Russel, Stuart Artificial Intelligence: A modern Approach, Prentice Hall, NJ, 1995
  4. Patterson, Dan W. Introduction to Artificial Intelligence and Expert Systems, Prentice Hall of India Private Limited New Delhi, 1998


#Things To Remember