The UGC NET Computer Science (Artificial Intelligence)

 The UGC NET Computer Science (Artificial Intelligence)

Introduction

In this article, you will know and learn about some important terminologies, and MCQs frequently asked in the UGC NET Computer Science.

Artificial Intelligence (AI)

·       John Mc Carthy coined the term AI in the 1950s.

·       AI is a branch of computer science.

·       AI is the simulation of human intelligence in machines programmed to think like humans and mimic their actions.

·       AI is used to create intelligent machines that can behave like humans, think like humans, and make decisions under different situations.

Artificial Intelligence (AI)

·       AI is a broad term referring to robots, bots, chatbots, androids, and cyborgs.

·       AI is a field of computer science that promotes the creation of intelligent machines that work and react like a human.

·       AI is a field of study that tries to make computers smart.

·       AI is the ability of a computer program or a machine to think and learn.

·       Some of the activities that computers with AI are designed for include speech recognition, learning and planning, and problem-solving.

·       The five basic components of AI are learning, reasoning, problem-solving, perception, and language understanding.

Click the given link to learn about Artificial Intelligence      AI-2

Machine Learning (ML)

·       ML is a branch of AI and computer science.

·       ML works on using data and algorithms to imitate how humans learn to improve the machine’s accuracy.

·       ML allows the user to feed a computer algorithm an immense amount of data and have the computer analyze and make data-driven recommendations and decisions based on only the input data.

·       ML is based on the idea that systems can learn from data, identify patterns, and make decisions with minimal human intervention.

·       ML allows systems to automatically learn and improve from experience without being explicitly programmed.

·       ML depends on mathematics and statistics. Understanding of some mathematical and statistical methods is required to modify ML models or to build/create new models.

Natural language processing (NLP)

·       NLP is a collective term referring to the automatic computational processing of human languages.

·       NLP is a subfield of linguistics, computer science, and AI concerned with the interactions between computers and human languages.

·       NLP refers to the branch of AI that gives computers the ability to understand text and spoken words in much the same way human beings can.

·       NLP helps a computer communicate with humans in their own language and scales other languages–related tasks.

·       NLP allows computers to read text, hear speech, interpret it, measure sentiment, and determine which parts are important.

Neural network

·       Neural network is also known as artificial neural network (ANN) or simulated neural network (SNN).

·       A neural network refers to a series of algorithms that endeavor to recognize underlying relationships in a data set through a process miming how the human brain operates.

·       Neural networks are networks used in ML that work like the human nervous system.

Supervised ML

·       Supervised ML is the ML task of learning a function that maps an input to an output based on example input-output pairs.

·       Supervised ML uses past data to make predictions.

·       Supervised ML uses classification and regression ML algorithms.

·       A common example of supervised ML is the spam filtering of e-mails.

Unsupervised ML

·       Unsupervised ML refers to a type of algorithm that learns patterns from untagged data.

·       Unsupervised ML deals with unlabelled data.

·       Unsupervised ML finds hidden patterns. Through mimicry, the machine is forced to build a compact internal representation of its world and generate imaginative content.

·       Unsupervised ML allows the model to work on its own to discover patterns and information that was previously undetected.

·       Unsupervised ML uses clustering and association ML algorithms.

·       A common example of unsupervised ML is Facebook.

Reinforcement ML

·       Reinforcement learning is a type of ML method where an intelligent agent (computer program) interacts with the environment and learns to act within that.

·       Reinforcement learning in ANN is a goal-directed computational approach where an agent learns to perform a task by interacting with an unknown dynamic environment.

·       Reinforcement learning is about an autonomous agent taking suitable actions to maximize rewards in a particular environment.

·       Reinforcement learning is an ML training method based on rewarding desired behaviors and/or punishing undesired ones. During learning, the learning algorithm updates the agent policy parameters.

·       Some common examples of NLP that use reinforcement learning are predictive text, text summarization, question answering, and machine translation.

·       By studying typical language pattern reinforcement learning agents can mimic and predict how people speak to each other every day.

·       Reinforcement ML is used for improving or increasing efficiency.

Perceptrons

Perceptrons are two-layer networks with one input and one output.

Single-Layer Perceptron (SLP)

·       SLP stands for single-layer perceptron.

·       SLP is the simplest type of ANN.

·       A SLP is a feed-forward network based on a threshold transfer function.

·       A SLP can only learn linear functions.

·       SLP can classify only linearly separable cases with a binary target (1,0).

Multilayer perceptron(MLP)

·       MLP can learn non-linear functions.

·       A MLP contains one or more hidden layers (apart from one input and one output layer).

Space

In an AI problem, space refers to the exhaustive collection of all conceivable states.

State

·       AI problems can be represented as a set of well-formed states.

·       A state can be an initial state or a goal state.

State space

·       State space is known as the set of all possible and known states of a system.

·       In state space, each unique point represents a state of the system. For example, take a pendulum moving into and fro motion. The state of such an idealized pendulum is represented by its angle and its angular velocity.

·       Four state space forms are the phase variable form (controller form), the observer form, the modal form, and the Jordan form.

State space representation

·       State space representation is a mathematical model of a physical system expressed as a function of input, output, and state variables related by first-order differential equations or difference equations.

·       The state of the system can be represented as a vector within that space.

State-space models

·       State-space models are models that use state variables to describe a system by a set of first-order differential equations or difference equations.

State space representation of the problem

State space representation of the problem refers to a set of all possible states for a given problem. For example, chess game, the initial position of all the pieces on a chess board defines the initial state. 

Knowledge representation in AI

·       Knowledge representation in AI is a study of how the beliefs, intentions, and judgments of an intelligent agent can be expressed suitably for automated reasoning.

·       Knowledge representation in AI describes the representation of knowledge.

·       The four main approaches to knowledge representation in AI are simple relational knowledge, inheritable knowledge, inferential knowledge, and procedural knowledge.

Semantic networks

·       Semantic networks are a logic-based formalism for knowledge representation.

·       A Semantic network is a method of knowledge representation that represents semantic relations between concepts using a directed or undirected graph consisting of vertices (indicating concepts) and edges (including relations).

·       Semantic networks are a way of representing relationships between objects and ideas.

·       For example, a network might tell a computer the relationship between different animals (a cat 1’s a mammal, and a cat has whispers).

Knowledge representation in logic

·       Knowledge representation in logic gives processable form to all the information that can be precisely expressed in any other language. Logic allows the expression of all the information that can be stored in computer memory.

·       Three modern roles for logic in AI, which are based on the theory of tractable Boolean circuits are a) logic as a basis for computation b) logic for learning from a combination of data and knowledge c) logic for reasoning about the behavior of ML systems.

·       A common example of logical representation is, we can represent “Richard is a king, Jack is a person, and all kings are persons”. Using the following predictive logic: Richard is a king.

Ontology-based Knowledge Representation

·       Ontology-based knowledge representation describes the individual instances and roles in the domain that are represented using unary and binary predicates.

·       Ontology-based Knowledge representation enables knowledge sharing, processing, reuse, capturing, and communication.

·       Ontology seeks the classification and explanation of entities.

Planning

·       Planning is the process of determining various actions that often lead to a solution.

·       Planning is considered the logical side of acting.

·       The entire process of the planning system includes missions, objectives, policies, procedures, programs, budgets, and strategies.

·       The three main components of planning are strategic thinking, long-range planning, and operational planning.

Planning in AI

· Planning in AI is about the decision-making tasks performed by robots or computer programs to achieve a specific goal.

·       The execution of planning is about choosing a sequence of activities with a high likelihood of completing the specific tasks.

Different types of planning in AI are

·       Classical planning

·       Reduction to other problems

·       Temporal planning

·       Probabilistic planning

·       Preference-based planning

·       Conditional planning

·       Contingent planning

·       Conformant planning

·       FSSP (Forward State Space Planning)

·       BSSP (Backward State Space Planning)

FSSP (Forward State Space Planning)

FSSP says that given an initial state in any domain, we perform some necessary actions and obtain a new state S (which also contains some new terms), called a progression. It continues until we reach the target position.

BSSP (Backward State Space Planning)

In BSSP, we move from the target state g to the sub-goal g, tracing the previous action to achieve that goal. This process is called regression (going back to the previous goal or sub-goal). These sub-goals should also be checked for consistency.

Goal Stack Planning (GSP) in AI

·       GSP is a method in which we work backward from the goal state to the initial state. We make use of a stack to hold these goals that need to be fulfilled as well as the actions that we need to perform for the same.

·       GSP is designed to handle problems having compound goals.

·       GSP utilizes STRIP as a formal language for specifying and manipulating the world with which it is working.

Turing test

·       The Turing test is named after Alan Turing.

·       Turing test is a method of inquiry in AI for determining whether or not a computer is capable of thinking like a human being.

·       A Turing test requires humans, interrogators, and a computer.

·       Turing machines provide a powerful computational model for solving problems in computer science and testing the limits of computation.

·       Turing machines are similar to finite automata/finite state machines but have the advantage of unlimited memory.

·       A common example of Turing is the interview process for Turing.

CAPTCHA

·       CAPTCHA stands for Completely Automated Public Turing Test.

·       CAPTCHA is an interactive feature added to web forms to distinguish whether a human or automated agent is using the form.

Rational agent approach

·       A rational approach is also called a rational choice theory or rational action theory or choice theory.

·       Acting rationally means acting to achieve one’s goals, given one’s beliefs.

Agent

·       Agent is just something that perceives and acts.

·       An agent is anything that can be viewed as perceiving its environment through sensors and acting upon the environment through effectors.

·       An AI system is composed of an agent and its environment.

·       In the AI system, our aim is to design agents.

·       The four main types of agents are Artists agents, sales agents, distributors, and licensing agents.

Rational Agent in AI

·       Rational Agent in AI refers to a theoretical entity that considers realistic models of how people think, with preferences for advantageous outcomes and an ability to learn.

·       A Rational Agent in AI could be anything that makes decisions, such as a person, firm, machine, or software.

·       A Rational Agent in AI carries out an action with the best outcome after considering past and current percepts.

·       A Rational Agent performs the actions that cause the agent to be most successful.

·       Rational Agent is a computer program that performs tasks based on pre-defined rules and procedures.

Intelligent agent

·       An intelligent agent is a system that can perceive its environment and take actions to achieve a specific goal.

·       An intelligent agent is an advanced computer system that can gather, analyze and respond to the data it collects from its surrounding environment.

·       An intelligent agent can be a self-driving car or a virtual personal assistant.

Script

A script is a structured representation describing a stereotyped sequence of events in a particular context.

Frames

·       A frame is also known as slot-filter knowledge representation in AI.

·       Frames are the AI data structure that divides knowledge into substructures by representing stereotyped situations.

·       A frame is a sequence of bits consisting of the frame coordination bits such as a frame check sequence and the packet payload.

·       Frame-based systems contain the ideas of classes, instances, and inheritance. For example, the class vehicle could be defined, along with the subclasses car and truck.

·       A frame consists of a collection of slots and slot values.

·       The two main categories of frame structures are braced frame structure and rigid frame structure.

·       Some common examples of frames are V.42 modem frames, Fibre channel frames, PPP (Point-to-point) frames, and Ethernet frames.

Linear planning

·       Linear planning is the planning or scheduling of project management tasks where distance is a significant factor in the project. Some common examples of projects include roads, rail, pipelines, and transmission lines.

·       Linear planning considers the time factors of a task as well as the location factors.

·       Linear models are supervised learning algorithms used for solving either classification or regression problems.

Non-linear planning

·       Non-linear planning is used to set a goal stack and is included in the search space of all possible sub-goal orderings.

·       Non-linear planning handles the goal interactions by the interleaving method.

·       Non-linear planning takes a larger search space since all possible goal orderings are considered.

·       A common example of non-linear planning in AI is “start the car and put on the Bluetooth, then complete task 2 of a phone call, and then finally, complete task 1 by leaving the car at the service station”.

Linear programming

Linear programming is a method to achieve the best outcome in a mathematical model whose requirements are represented by linear relationships.

Non-linear programming

Non-linear programming is a process of solving an optimization problem where the constraints or the objective functions are non-linear.