18CSC305J - Artificial Intelligence UNIT 1

 12M:

Explain the types of intelligent agents.

Agent:

  • An agent is anything is anything that can be viewed as perceiving its environment through sensors and acting upon the environment through actuators

Agent Program:

  • Agent’s behavior is mathematically described by an agent function

  • Agent’s behavior is practically described by an agent program

  • It is the real implementation

  • AI designs the agent program

  • To design an agent program, we need to understand

  • Percepts

  • Actions

  • Goals

  • Environment 

Types:

  • Simple reflex agents

  • Model based reflex agents

  • Goal based agents

  • Utility based agents


Simple reflex agents:

  • Uses just condition action rules

  • Works only if the environment is fully observable


Model based reflex agents:

  • For the world that is partially observable

  • The agent has to keep track of an internal state that depends on the percept history


Goal based agents:

  • Current state of the environment is always not enough

  • The goal is to achieve correctness


Utility based agents:

  • Goals alone are not enough to generate high quality behavior

  • If goal means success, utility means the degree of success



Problem Solving: 

Problem solving process:

Components of problems:

  • A problem can be defined formally by 5 components:

  • Initial state

  • actions/ operators

  • Transition model

  • A goal test

  • Path cost

Problem formulation:

  • Choosing relevant set of states and feasible set of operators for moving from one state to another

Toy world problem:

  • Intended to illustrate various problem solving methods

Real world problems:

  • One whose solutions people actually care about

Problem types:

  • Deterministic or observable

  • Non observable

  • Non deterministic or partially observable

  • Unknown state space

Example: (Toy world problem)

4-Queens: 

Given a 4 x 4 chessboard, we have to place 4 queens such that no two queens attack each other



4M:


What is an agent? / What is an agent program?


Agent:

  • An agent is anything is anything that can be viewed as perceiving its environment through sensors and acting upon the environment through actuators

Agent Program:

  • Agent’s behavior is mathematically described by an agent function

  • Agent’s behavior is practically described by an agent program

  • It is the real implementation

  • AI designs the agent program

  • To design an agent program, we need to understand

  • Percepts

  • Actions

  • Goals

  • Environment 



What is meant by a cognitive model?

  • The goal is to make systems think like humans

  • Simulating human problem solving and mental processing in a computerized model

  • It is used in various AI applications like NLP, robotics, VR, neural networks



Give complete solutions to 4 Queens problem

  • The aim of the n queens problem is to place n queens in a chessboard in such a way that no queens can attack each other ( no two queens are placed diagonally, vertically or horizontally)



What are the important AI techniques?

  • AI technique is a method that achieves knowledge.

  • The main AI techniques are

  • Search

  • Use of knowledge

  • Abstraction 

  • A search program finds solution for a problem by trying various sequences of actions until a solution is found (best way but not always possible because of large state space)

  • The use of knowledge provides a way of solving complicated problems by manipulating the structures of the objects that are concerned

  • Abstraction finds a way of separating important features and notifications from the unimportant ones that would otherwise confuse any process



Types of constraints in a CSP problem

  • CSP is a special class of search problems

  • The goal test is a set of constraints

  • Types of constraints:

  • Unary constraints: involve a single variable (Eg: SA != green)

  • Binary constraints: involve pairs of variables (Eg: SA != WA)

  • Higher order constraints: involve 3 or more variables (Eg: O + O = R + 10 . X1)

  • Preferences (soft constraints): gives constrained optimization problems (Eg: red is better than green)



What are the various constraint satisfaction problems?

  • Finding a solution that meets a set of constraints is the goal of constraint satisfaction problems

  • There are mainly 3 basic components of a CSP:

  • Variables

  • Domains

  • Constraints 

  • CSP can be categorized according to the type of domain:

  • Finite: finite domains have a finite number of possible values such as colors or integers

  • Infinite: infinite domains have a infinite number of possible values such as real numbers

  • Continuous: continuous domains have a infinite number of possible values but can be represented by a finite set of parameters

  • Examples:

  • Cryptarithmetic puzzles

  • Map coloring problems

  • Sudoku

  • crosswords



Give PEAS description for online bus reservation system

PEAS stands for 

  • Performance measures

  • Environment

  • Actuators 

  • Sensors 

Performance measures: 

  • Response time

  • Accuracy

  • User satisfaction

Environment:

  • Users

  • Bus routes

  • Payment gateways

Actuators: 

  • Seat reservation

  • Cancellation

  • Payment processing

Sensors:

  • Seat availability sensors

  • Payment status sensors



Describe various AI models


What are the characteristics of intelligent agents?

  • An agent is anything that can be viewed as perceiving its environment through sensors and acting upon that environment through actuators

  • Common characteristics of intelligent agents are adaptation based on experience, real-time problem-solving, analysis of error or success rates, and the use of memory-based storage and retrieval.

  • Human agent: eyes, ears, and other organs for sensors; hands,legs, mouth, and other body parts for actuators

  • Robotic agent: cameras and infrared range finders for sensors; various motors for actuators



What are problem solving agents?

  • A simple problem solving agent first 

  • formulates a goal and a problem

  • Searches for a sequence of actions that would solve the problem

  • Then execute the actions one at a time

  • When this is complete, it formulates another goal and starts over



What are the problems in simple reflex agents?

  • They cannot compute complex equations or solve complicated problems. 

  • They work only in environments that are fully-observable in the current percept, ignoring any percept history.



What are the conceptual components of a Learning agent?    

Four conceptual components

  • Learning element - Making improvement

  • Performance element - Selecting external actions

  • Critic - Tells the Learning element how well the agent is doing with respect to fixed performance standards. (Feedback from users or examples, good or not?)

  • Problem generator - Suggest actions that will lead to new and informative experiences.



What are the parameters of   Measuring problem-solving performance?


  • the search cost--how long the agent takes to come up with the solution to the problem, and

  • the path cost--how expensive the actions in the solution are.

The total cost of the solution is the sum of the above two quantities.


  • Completeness: Is the algorithm guaranteed to find  a solution when there is one?

  • Optimality: Does the strategy find the optimal  solution (i.e., lowest path cost among all solutions)

  • Time complexity: How long does it take to find a  solution?

  • Space complexity: How much memory is needed  to perform the search?



Define heuristic function

  • Backtracking allows to go to the previous decision-making  node to eliminate the invalid search space with respect to  constraints

  • Heuristics plays a very important role here

  • If we are in position to determine which variables should be  assigned next, then backtracking can be improved

  • Heuristics help in deciding the initial state as well as  subsequent selected states

  • Selection of a variable with minimum number of possible  values can help in simplifying the search

  • This is called as Minimum Remaining Values Heuristic (MRV)  or Most Constraint Variable Heuristic



Data acquisition:

  • There are 2 things that AI needs code and training data

  • The more accurate the training data, the more powerful the tool



What are rational agents?

  • Ideal Rational agents are ones that are capable of doing expected actions to maximize its performance measure

  • Rationality of an agent depends on

  • Performance measure

  • Percept sequence 

  • Knowledge about environment

  • Action able to be performed

  • The problem can be categorized by PEAS

Comments

Post a Comment

Popular posts from this blog

18ECO124T - HUMAN ASSIST DEVICES UNIT 4 & 5 - 12M

Image
12M: SET 1: Write in detail about the pediatric orthosis with a diagram Pediatric orthosis: There are several respects in which a child's orthotic requirements differ from those of an adult and these may extend beyond treatment of the obvious abnormality. Accordingly, before considering specific hardware and research activities, the criteria affecting the performance of an orthosis will be discussed. A child is a particularly dynamic person in terms both of his activity and of his developing personality.  It is important that there is a full appreciation of these factors so that prescription of an orthosis can be made sympathetically and in harmony with his development Human factors: Many physical and emotional factors can affect the success of an orthosis Physical factors can be considered under the headings of deformity, growth and activity Deformity:  Frequently an orthosis is fitted to prevent the development of a deformity, particularly during periods of rapid growth For ...
Read more

18CSE359T - NATURAL LANGUAGE PROCESSING UNIT 1 - 12M

Image
  12M: Using the different NLP methods, identify POS tagging The POS tagging problem is to determine the POS tag for a particular istance of a word POS - Parts of Speech Methods to assign POS to words: Rule based tagging Learning based: Stochastic models: HMM tagging, maximum entropy tagging Rule learning: transformation based tagging Rule based tagging: Two stage architectureENGTWOL tagging Stage 1 - Use dictionary to assign each word a list of potential POS Stage 2 - Hand-written disambiguation rules to fine tune to single POS Disambiguation rule: Adverbial-THAT rule:- Given input: ‘that’ if   (+1 A/ADV/QUANT);  (+2SENT-LIM);  (NOT -1 SVOC/A); then eliminate non-ADV tags else eliminate ADV tag I consider that odd. NOT ADV It isn’t that odd. ADV Stochastic POS Tagging: Pick the most likely tag for the word Example:  She is expected to race tomorrow The race for the cup begins! Transformation based tagging: Combination of rule based tagging and stochast...
Read more

18CSE483T - INTELLIGENT MACHINING UNIT 4 & 5 - 4M

Image
  4M: Compare and contrast different Neural Network Architecture Feedforward NN Feedback NN Self-organizing NN Layers of neurons connected in a forward direction without cycles Layers can form cycles, allowing feedback loops Neurons are organized into a lattice structure Simplicity, ease of implementation Capturing temporal dependencies, dynamic behavior Adaptive learning, ability to detect patterns Suitable for structured data Suitable for temporal or sequential data Suitable for clustering and unsupervised learning Cannot capture temporal dependencies Complexity in training, potential instability Lack of supervised training, interpretability challenges Regression, classification tasks Time-series prediction, dynamic systems Pattern recognition, data visualization Compare and contrast forward propagation and backward propagation Forward propagation Backward propagation Calculate the output of the neural network Adjust weights to minimize the error Input data is passed through the ...
Read more