Conventional space wastage and numeric processing will
Conventional Vs Intelligent Computing
Conventional Computing is a technique to execute what it is programmed to do.
1. It is necessary to define all the possible ways to solve a problem.
2. Rules are coded hence are incomprehensible (usually difficult).
3. Explanation and reasoning are not provided.
4. Algorithms are used for searching.
5. Sequential processing takes place.
Intelligent Computing is a technique which is capable of self-learning and emulates an intelligent outcome.
1. Only knowledge (logic) needs to be defined.
2. Written in everyday language so they are easy to understand.
3. Explanation reasoning is given.
4. Heuristic searching is used.
5. Multitasking takes place.
6. Used to solve complex problems.
Example: In case of Tic-tac-toe program can be solved via both conventional computing as well as intelligent computing but if solved this problem using–
a) Conventional computing: It will be time-consuming, space wastage and numeric processing will take place as it looks at the squares one at a time. Valid reasoning will not be provided.
Hence, it will be difficult to understand.
b) Intelligent computing: Knowledge is used to solve the problem by exploiting the structures of the objects involved and this approach solves the problem similarly as human beings do. People act as parallel processors and look at several parts of the board at once.
Agents are functional systems which decide independently which action to take in a given situation.
It has goals, sensors and actuators.
A) Learning Agent
– Has performance element that decides which actions to take.
– Has learning element which makes the performance element more efficient.
– Critic component which evaluates the efficiency of an agent and provides feedback.
– Has a problem generator.
B) Utility Agent
– It maps a state to a real number, checks how efficient agent is in achieving the goal.
Conclusion: Intelligent computing is better than conventional computing as it involves perceiving, understanding the problem then reasoning, learning and adapting the solutions these techniques will work for large problems where usually direct methods won’t work. This increases the efficiency of computation of programs.