RV Institute of Technology and

Management®

Rashtreeya Sikshana Samithi Trust

RV Institute of Technology and Management ®
(Affiliated to VTU, Belagavi)
JP Nagar, Bengaluru – 560076

Department of Computer Science and Engineering

Course Name: Introduction to AI and its Applications

Course Code: 1BAIA103/203
I Semester 2025 Scheme
Prepared By:
1. Prof Padmasree N
Assistant Professor, Department of Computer Science and
Engineering, RVITM, Bengaluru – 560076
Email: [email protected]

2. Prof. Uppin Rashmi

Assistant Professor, Department of Computer Science and
Engineering, RVITM, Bengaluru – 560076
Email: [email protected]

3. Prof. Manjusha Kulkarni

Assistant Professor, Department of Computer Science and
Engineering, RVITM, Bengaluru – 560076
Email: [email protected]
 RV Institute of Technology and
Management®

Chapter 1

Introduction to Artificial Intelligence

1.1 What is Artificial Intelligence?

• Definition: Artificial intelligence (AI) is the science and engineering of making intelligent machines,
particularly intelligent computer programs.
• Layman's View: AI means intelligence demonstrated by machines that mimic human actions, learn
from experiences, adjust to new inputs, and perform human-like tasks.
• Researcher's View: AI is a set of algorithms that generates results without explicit instructions, making
machines capable of thinking and acting rationally and humanely.
• Core Tasks: AI applications perform specialized tasks by processing large amounts of data and
recognizing patterns. They can learn from experience, recognize objects, understand/respond to language,
and make decisions to solve real-world problems.
• Goal: To build machines and algorithms capable of performing computational tasks that typically
require human-like brain functions.
• NITI Aayog Definition: AI refers to the ability of machines to perform cognitive tasks like thinking,
perceiving, learning, problem-solving, and decision-making. It has evolved beyond mimicking human
intelligence, enabling intelligent systems to take over tasks, enhance connectivity, and improve
productivity.

1.1.1 How Does AI Work?

• AI systems operate effectively when given large amounts of labelled training data5.
• This data is analysed to discover correlations and patterns, which are then used to make predictions
about future states.
• Example: A chatbot learns to converse with humans by being fed examples of text chats.
• AI programming focuses on three cognitive skills5
◦ Learning Processes: Acquiring data and creating rules (algorithms) to turn data into actionable
information.
◦ Reasoning Processes: Choosing the right algorithm for the desired outcome6.
◦ Self-Correction Processes: Continuously enhancing algorithms for accurate results6.
• Importance: AI allows businesses to gain insights, automate redundant jobs, reduce costs, and increase
revenue....

1.1.2 Advantages and Disadvantages of Artificial Intelligence

 RV Institute of Technology and
Management®

• Advantages:
◦ Performs well on tasks using detailed data.
◦ Takes less time to process huge data volumes.
◦ Generates consistent and accurate results.
◦ Can be used 24x79.
◦ Optimises tasks by better utilising resources.
◦ Automates complex processes.
◦ Minimises downtime by predicting maintenance needs.
◦ Enables new, better quality, and faster product production.
• Disadvantages:
◦ Involves more cost.
◦ Requires technical expertise to develop and use AI applications.
◦ Lack of trained professionals.
◦ Incomplete or inaccurate data can lead to disastrous results.
◦ Lacks the capability to generalise tasks.
1.2 History of Artificial Intelligence
• 1943: Warren McCullough and Walter Pitts proposed the first mathematical model for a neural network.
• 1950:
◦ Alan Turing demonstrated the Turing Test.
◦ Marvin Minsky and Dean Edmonds built the first neural network computer.
◦ Claude Shannon published on programming a computer for chess.
◦ Isaac Asimov published "Three Laws of Robotics".
• 1952: Arthur Samuel developed a self-learning program to play checkers
• 1954: IBM computer translated 60 Russian sentences into English.
• 1956: John McCarthy coined the term ‘artificial intelligence’ and discussions at a conference marked
AI's birth.
• 1958: John McCarthy developed the AI programming language Lisp.
• 1959: Allen Newell, Herbert Simon, and J.C. Shaw developed the General Problem Solver (GPS);
Arthur Samuel coined ‘machine learning’.
• 1963: John McCarthy started the AI Lab at Stanford.
• 1966: Joseph Weizenbaum developed ELIZA, an early natural language processing program.
• 1969: First successful expert system for diagnosing blood infections developed at Stanford.
• 1972: Logic programming language PROLOG was created.
• 1974-1980: The 'First AI Winter' due to DARPA funding cutbacks.
• 1980: DEC developed R1, the first successful commercial expert system.
• 1982: Japan launched the Fifth Generation Computer Systems (FGCS) project.
• 2008: Google introduced speech recognition in its iPhone app.
• 2011: Apple released Siri, an AI-powered virtual assistant
• 2012: Andrew Ng's Google Brain project enabled a neural network to recognise a cat from million
YouTube videos.
 RV Institute of Technology and
Management®

• 2014: Google's first self-driving car passed a state driving test; Amazon released Alexa.
• 2015: Baidu’s Minwa supercomputer achieved higher image identification accuracy than humans using a
convolutional neural network.
• 2016: Google DeepMind’s AlphaGo defeated world champion Go player Lee Sedol; Sophia, the first
'robot citizen', was created.
• 2018: Google released NLP engine BERT; Waymo One self-driving service launched.
• 2020: Baidu released its LinearFold AI algorithm for SARS-CoV-2 vaccine development.
1.3 Types of Artificial Intelligence AI systems can be classified into two main categories and four
functional groups1819:
• Weak AI (Narrow AI)1820:
◦ Designed to perform a specific task.
◦ Examples: Siri, Alexa, weather prediction, stock price optimisation, Google search, image
recognition, self-driving cars, IBM Watson1820.
◦ Operates within a limited context and is the most successful AI realisation to date.
• Strong AI (Artificial General Intelligence - AGI, or Artificial Super Intelligence - ASI)2021:
◦ Aims to resemble the human brain, utilising cognitive skills and fuzzy logic to perform tasks for
which it was not explicitly trained.
◦ Requires capabilities like visual perception, speech recognition, decision-making, and language
translation.
◦ Currently seen only in sci-fi movies, but believed that ASI will surpass human intelligence
• Four Groups Based on Functionality:
◦ Reactive Machines1922:
▪ Very basic, with no memory to store past experiences for future actions.
▪ Perceive the world and react to it.
▪ Example: IBM’s Deep Blue (defeated chess grandmaster Kasparov).
▪ Cannot improve with practice19.
▪ Limited number of specialized tasks, trustworthy, and reliable.
▪ Google's AlphaGo is also a game-playing reactive machine, but evaluates future moves using
neural networks.
◦ Limited Memory...:
▪ Retain data for a short period and use it for a limited time.
▪ Cannot permanently add data to an experience library.
▪ Used in autonomous vehicles (e.g., storing recent speed of nearby cars, distance between cars,
speed limits).
▪ More complex than reactive machines, continuously train models, and improve with feedback.
▪ Major machine learning models applying limited memory AI:
• Reinforcement Learning: Continuous learning via trial-and-error for better predictions.
• Long Short-term Memory (LSTM): Uses past data to predict the next item in a sequence,
prioritising recent information25.
• Evolutionary Generative Adversarial Networks (E-GAN): Evolves to explore new ways of
 RV Institute of Technology and
Management®

utilising past experiences for new decisions, using simulations and statistics.
◦ Theory of Mind:
▪ Focuses on imitating the human brain by forming representations about the world, including
thoughts, emotions, and memories26.
▪ Currently theoretical, but may become reality26.
▪ Machines would make decisions considering feelings from self-reflection and determination.
◦ Self-Awareness:
▪ The next step after Theory of Mind, incorporating human-level consciousness27.
▪ Systems would understand their own existence and use that information to deduce others' feelings.
▪ Could interpret user's feelings from both explicit communication and manner of communication.
▪ Knowledge of conscious context would enable responses to events.
▪ Does not yet exist.
1.4 Is Artificial Intelligence Same as Augmented Intelligence and Cognitive Computing?
• Augmented Intelligence:
◦ Weak AI that simply improves products and services.
◦ Example: Automatically highlighting vital information in business.
• Artificial Intelligence (True AI/Strong AI/AGI):
◦ The future AI that would far surpass the human brain's ability.
◦ Currently largely in the realm of science fiction, but technologies like quantum computing could make
it a reality.
◦ In reference to machines, it involves simulating how humans sense, learn, process, and react to
information.
• Cognitive Computing:
◦ Term for products and services that mimic and augment human thought processes.
1.5 Machine Learning and Deep Learning
• Machine Learning (ML):
◦ A branch of computer science that analyses data and identifies patterns to teach a machine to deduce
results and make decisions without human intervention.
◦ ML algorithms learn from experiences rather than explicit instructions.
◦ Automatically learn and improve by analysing datasets and comparing output, repeating the process
until accuracy improves.
◦ Enables machines to make data-driven decisions rather than being explicitly programmed for every
task.
◦ Relationship with AI: ML is an application and a subset of AI... AI is the superset, and ML is a
way to achieve AI.
• Traditional Programming vs. Machine Learning...:
◦ Traditional Programming: Manually creating a program with explicit rules/code in procedural
languages (e.g., C, Java, Python) that accepts data and returns output.
◦ Machine Learning: Automated process where algorithms automatically formulate rules from data,
adding embedded analytics (e.g., natural language interfaces, outlier detection, recommendations)35. Uses
 RV Institute of Technology and
Management®

pre-written algorithms that learn how to solve problems themselves.

◦ Capabilities: Different, but ML supplements conventional programming (e.g., ML for predictive
algorithms, traditional for UI design or data visualisation).
• Deep Learning (DL):
◦ An advanced machine learning technique that processes data inputs through multiple layers of
biologically-inspired neural networks.
◦ These hidden layers allow machines to learn "deeply," making connections and weighting inputs for
best results.
◦ Relationship with ML: DL is an advanced ML technique and a subset of ML3940.
• How AI Works (features)...:
◦ Autonomous: Makes independent decisions without human intervention, learning through input data
and past experiences.
◦ Predict and Adapt: Understands data patterns for decisions and predictions.
◦ Continuously Learns: Learns from patterns in data.
◦ Reactive: Perceives a problem and acts on perception
◦ Data Driven: Rise of data-centric AI systems due to cheaper data storage, fast processors, and
sophisticated deep learning algorithms.
◦ Accurate Predictions: Outperformed humans based on past experiences, success depends on
correctly labelled large datasets.
◦ Futuristic: Scope continuously expanding.

Chapter 3: Artificially Intelligent Machine

3.1 Defining Intelligence
• Howard Gardner's Categories of Intelligence...:
◦ Linguistic intelligence: Ability to speak, recognise, and use phonology, syntax, and semantics
(narrators, orators).
◦ Musical intelligence: Ability to create, communicate with, and understand sounds, pitch, rhythm
(musicians, composers).
◦ Logical-mathematical intelligence: Ability to use and understand complex, abstract ideas
(mathematicians, scientists).
◦ Spatial intelligence: Ability to perceive visual/spatial information, change, and recreate visual images
(map readers, astronauts, physicists)
◦ Bodily-kinesthetic intelligence: Ability to use the body to solve problems or manipulate objects
(players, dancers).
◦ Intrapersonal intelligence: Ability to distinguish one's own feelings, intentions, and motivations
(Gautam Buddha).
◦ Interpersonal intelligence: Ability to recognise and differentiate others' feelings, beliefs, and
intentions (mass communicators, interviewers).
• A machine is artificially intelligent if it exhibits at least one, and at most all, of these intelligences.
 RV Institute of Technology and
Management®

3.2 Components of Intelligence Intelligence is composed of:

• Reasoning: Processes used in making decisions and predictions. Two types: Inductive and
Deductive4849.
◦ Inductive Reasoning: Uses specific observations to make broad general statements; conclusion is
likely but not certain.
◦ Deductive Reasoning: Uses general premises to reach specific, certain conclusions49.
• Learning: Gaining knowledge or skill through studying, practicing, being taught, or experiencing49.
Ability to improve awareness
◦ Auditory learning: By listening.
◦ Episodic learning: Remembering sequences of events.
◦ Motor learning: By precise muscle movement.
◦ Observational learning: By watching and imitating.
◦ Perceptual learning: By recognising stimuli seen before.
◦ Relational learning: Differentiating stimuli based on relational properties.
◦ Spatial learning: Through visual stimuli.
◦ Stimulus-response learning: Performing behaviour when stimulus is received.
• Problem Solving: Finding a desired solution when the path is blocked by hurdles, using decision-
making.
• Perception: Acquiring, interpreting, selecting, and organizing sensory information. Humans use sensory
organs, AI uses data from sensors.
• Linguistic intelligence: Ability to use, comprehend, speak, and write verbal/written language in
interpersonal communication.
3.3 Differences Between Human and Machine Intelligence
• Perception: Humans perceive by patterns; machines analyse data with respect to rules.
• Information Storage/Recall: Humans use patterns; machines use searching algorithms.
• Handling Missing/Distorted Information: Humans can deduce missing/distorted info; machines lack
this ability with high accuracy.
3.4 Agent and Environment
• An AI system comprises an agent and its environment.
• Agent: Anything that makes decisions and is capable of perceiving its environment (e.g., a person, firm,
machine, software).
• Sensors: Help agents perceive their environment.
• Effectors: Help agents act upon their environment.
• Types of Agents:
◦ Human agent: Sensory organs as sensors, hands/legs/mouth as effectors.
◦ Robotic agent: Cameras/infrared range finders as sensors, motors/actuators as effectors.
◦ Software agent: Bit strings as programs and actions.
3.4.1 Key Terminology
• Performance Measure of Agent: Criteria determining agent's success.
• Behaviour of Agent: Action performed after receiving a percept.
 RV Institute of Technology and
Management®

• Percept: Perceptual inputs given to an agent at a specific instance.

• Percept Sequence: List of all percepts received by an agent till date.
• Agent Function: A map from the percept sequence to an action.
3.4.2 Rationality
• Rationality: Inculcates responsibility, sensibility, and judgment, empowering the agent to perform
expected actions after perceiving.
• Dependence: Rationality depends on:
◦ Agent’s performance measure.
◦ Agent’s percept sequence so far.
◦ Agent’s prior knowledge about the environment.
◦ Agent’s possible actions.
• A rational agent performs actions to maximise its performance.
• A problem solved by an agent is characterised by Performance, Environment, Actuators, and Sensors
(PEAS).
3.4.3 Structure of Intelligent Agents
• Agent = Architecture + Agent Program.
◦ Architecture: The machinery on which an agent works.
◦ Agent Program: Implementation of an agent function.
3.4.4 Types of Agents
• Simple Reflex Agents
◦ Choose actions based only on the current percept.
◦ Rational only if environment is completely observable.
◦ Work using condition-action rules (If condition then action).
◦ Problems: Limited intelligence, no knowledge of past states, updates required for environment
changes, can get stuck in infinite loops in partially observable environments.
• Model-Based Reflex Agents:
◦ Use a model of the world to choose actions, maintaining an internal state.
◦ Model: Knowledge about "how things happen in the world".
◦ Internal state represents unobserved aspects of the current state based on percept history.
◦ Requires information on how the world evolves and how agent actions affect it.
• Goal-Based Agents:
◦ Choose actions to achieve specific goals.
◦ Offers more flexibility than reflex agents as knowledge is explicitly modelled and modifiable.
• Utility-Based Agents:
◦ Used when goals conflict or are difficult to achieve.
◦ Choose actions based on a preference (utility) for each state, maximising expected utility or
'happiness'.
• Learning Agent...:
◦ Learns from past experiences.
◦ Starts with basic knowledge and adapts automatically through learning.
 RV Institute of Technology and
Management®

◦ Four main conceptual components:

▪ Learning element: Responsible for improvements by learning from the environment.
▪ Critic: Provides feedback to the learning element, evaluating performance against a standard.
▪ Performance element: Chooses external action.
▪ Problem generator: Suggests actions for new and informative experiences.
3.4.5 The Nature of Environments
• AI programs can operate in confined or unlimited domains.
• Softbots: Software agents operating in detailed, complex environments, choosing actions in real-time.
• Turing Test Environment: Artificial agents tested on equal ground with real agents to determine
intelligent behaviour.
◦ Turing Test: A human interrogator interacts via typing with a human and a machine in separate
rooms. If the interrogator cannot distinguish the machine from the human, the machine is considered
intelligent.
3.4.6 Types of Environments Environments can be categorised by dimensions:
• Discrete/Continuous:
◦ Discrete: Limited, distinct, clearly defined states (e.g., chess game).
◦ Continuous: Infinite or large number of states (e.g., self-driving car).
• Observable/Partially Observable/Unobservable:
◦ Observable: Agent can determine the complete state from percepts at each time point.
◦ Partially Observable: Agent cannot determine complete state (e.g., due to noise, inaccuracy, missing
data, or task framework).
◦ Unobservable: Agent has no sensors.
◦ Fully observable environments do not need to maintain internal state; partially observable ones do.
◦ Example: Classic chess is fully observable; Kriegspiel chess is partially observable68.
• Accessible/Inaccessible:
◦ Accessible: Agent's sensory apparatus has access to complete state of environment (e.g., an empty
room's temperature)
◦ Inaccessible: Complete and accurate information about state is not obtainable (e.g., Earth event
information)
• Episodic/Non-episodic (Sequential):
◦ Episodic: Each episode involves perceiving and acting; action quality depends only on that episode;
subsequent episodes are independent. Simpler as agent doesn't need to think ahead.
◦ Non-episodic (Sequential): Agent requires memory to store past actions to determine next best
actions; current decisions affect future decisions.
3.5 Search
• AI agents use search algorithms to achieve tasks, especially in single-player games (e.g., Sudoku,
crossword).
• A search problem consists of:
◦ State space: Set of all possible states an agent can attain.
◦ Start state: Where searching begins.
 RV Institute of Technology and
Management®

◦ Goal test: Function to check if current state is the goal state.

◦ Solution: Sequence of actions (plan) transforming start to goal state, realised by search algorithms.
3.5.1 Types of Search Algorithms
• Categorised as informed and uninformed.
3.5.2 Properties of Search Algorithms
• Completeness: Returns at least one solution if one exists.
• Optimality: Returns the best solution (lowest path cost).
• Time and space complexity: Time taken and maximum storage required.
3.6 Uninformed Search Algorithms (Blind Search)
• Have no additional information about the goal state beyond the problem definition.
• Only know how to traverse or visit nodes.
• Blindly follow techniques regardless of efficiency.
• Information for each algorithm: Problem graph, Strategy, Fringe (data structure for possible states),
Tree, Solution plan.
3.6.1 Depth First Search (DFS)
• Strategy: Explores as far as possible along each branch before backtracking.
• Working: Starts from root, traverses deepest node first.
• Implementation: Uses a stack data structure.
• Advantages: Memory efficient, faster execution, terminates in finite time.
• Disadvantages: Incomplete (may not find solution even if it exists due to limit constraint), not optimal if
multiple solutions exist (may not find the best one), cannot check duplicate nodes.
• Complexity: Time and space complexity depend on path length and depth.
3.6.2 Depth-Limited Search Algorithm (DLS)
• Extension of DFS: Introduces a depth limit (ℓ) to prevent infinite loops.
• Working: Nodes at the depth limit are treated as leaf nodes.
• Advantages: Memory efficient, faster execution, terminates in finite time.
• Disadvantages: Incomplete (if solution is beyond limit), not optimal (may not find best solution even if
ℓ > d).
3.6.3 Breadth First Search (BFS)
• Strategy: Traverses the graph layer-wise (breadth-wise)....
• Working: Starts from root, explores all neighbour nodes at the present depth before moving to the next
level.
• Implementation: Uses a FIFO queue.
• Advantages: Finds a solution if it exists, finds the optimal solution if step cost is uniform8182.
• Disadvantages: High time and space complexity.
• Applications: Crawlers in Search Engines, GPS Navigation Systems, Finding Shortest Path for
Unweighted Graphs
3.6.4 Uniform Cost Search (UCS)
• Strategy: Finds an optimal solution when step costs are not the same
• Working: Computes the cumulative cost to expand each node from root to goal.
 RV Institute of Technology and
Management®

• Implementation: Uses a priority queue.

• Advantages: Finds optimal solution (by least cost), complete if states are finite and no zero-weight
loops, optimal if no negative cost.
• Disadvantages: May get stuck in infinite loops (considers only cost, not number of steps)84.
• Equivalence to BFS: UCS is equivalent to BFS if all path costs are the same85.
3.6.5 Iterative Deepening Depth-First Search (IDDFS)
• Combination: Combines DFS (memory efficiency) and BFS (fast search)8687.
• Working: Gradually increases the depth limit for DLS until the goal is found8788.
• Advantages: Combines benefits of BFS and DFS, complete if branching factor is finite, optimal if path
cost is non-decreasing86.
• Applications: Widely used when search space is large and goal node depth is unknown86.
3.6.6 Bidirectional Search
• Strategy: Searches simultaneously from both initial and goal states until they meet8789.
3.7 Informed Search Algorithms (Heuristic Search)
• Contain information about how far from the goal, path cost, etc.90.
• This knowledge helps agents explore less search space and reach the goal more efficiently90.
• Uses a heuristic function that estimates how close the agent is from the goal90.
• May not always give the best solution but guarantees a good solution in reasonable time90.
3.7.2 Best-First Search Algorithm (Greedy Search)
• Combination: Combines DFS and BFS91.
• Strategy: Always selects the path that appears best at that moment91.
• Working: Uses the heuristic function to choose the most promising node (closest to the goal)91.
• Implementation: Uses a priority queue91.

--------------------------------------------------------------------------------

Chapter 4: Knowledge Representation

4.1 Introduction
• To build AI systems with conscience, knowledge needs to be inculcated in them92.
• Knowledge and Intelligence: Knowledge of the real world is crucial for intelligence and creating
intelligent AI agents92. Without knowledge, decision-makers cannot sense the environment accurately or
make appropriate decisions93.
4.2 Knowledge Representation (KR)
• Focus: How AI agents think and how their thinking contributes to intelligent behaviour94.
• Purpose: To represent real-world information in a computer-understandable way so AI can solve
complex problems (e.g., medical diagnosis, natural language communication)94.
• KR involves modelling intelligent behaviour for an agent by representing beliefs, intentions, and
judgments95.
• It goes beyond merely storing data; it facilitates machines to learn from knowledge and experiences to
behave intelligently95.
 RV Institute of Technology and
Management®

4.2.1 What Knowledge Needs to be Represented? To make an AI system truly intelligent, it needs to
incorporate human-like intuition, intentions, prejudices, beliefs, judgments, common sense, and facts96:
• Object: Information and facts about relevant objects (e.g., vehicles, roads in a self-driving car)96.
• Events: Information and facts about actions occurring in the real world (e.g., applying brakes)97.
• Performance: How actions are performed (behaviour)97.
• Meta-knowledge: Knowledge about knowledge (what we know)97.
• Facts: Truths about the real world97.
• Knowledge base: Stores a group of technical sentences97.
4.2.2 What is Knowledge? Knowledge is the basic element for logical understanding, gained by
experience98. Five types of knowledge:
• Meta knowledge: Knowledge about knowledge98.
• Heuristic knowledge: Knowledge about a specific topic, often from experts, based on experience (rule
of thumb)98.
• Procedural knowledge (imperative knowledge): Information on "how to" achieve something,
including rules, strategies, procedures99.
• Declarative knowledge: Information about an object, describing concepts, facts, and attributes
(descriptive knowledge)99.
• Structural knowledge: Basic knowledge for complex problems, describing relationships between
concepts/objects (e.g., "kind of," "part of")100.
4.2.3 What is Logic?
• Logic: Main component of knowledge, facilitating drawing conclusions by filtering information100.
• In AI, knowledge is represented using logic, which has three main elements100:
◦ Syntax: Rules specifying how legal sentences in a language are constructed100.
◦ Semantics: Defines the meaning of syntactically correct sentences, relating to the real world101.
◦ Logical Inference: Deducing conclusions from facts/problems using inference algorithms101.
4.2.4 Cycle of Knowledge Representation in AI An AI system's intelligent behaviour involves several
components102:
• Perception: Retrieves data from the environment using sensors, identifies noise sources, checks damage,
defines responses102.
• Learning: Learns from data captured by perception, focusing on self-improvement through knowledge
acquisition, inference, and heuristics103.
• Knowledge Representation and Reasoning (KRR): The core component for human-like intelligence.
Defines what an agent needs to know and how automated reasoning procedures make this knowledge
available103.
• Planning and Execution: Analyses KRR. Planning selects an initial state, enumerates
preconditions/effects, and sequences actions to achieve goals. Execution performs these actions104.
4.2.5 Knowledge Representation Requirements A good KR system must have104105:
• Representational accuracy: Represents all required knowledge104.
• Inferential adequacy: Manipulates structures to produce new knowledge105.
• Inferential efficiency: Directs inference mechanism to generate appropriate results105.
 RV Institute of Technology and
Management®

• Acquisitional efficiency: Easily acquires new knowledge automatically105.

4.3 Knowledge-Based Agent (KBA)
• Agents that mimic human knowledge, using knowledge and reasoning to act efficiently and make
appropriate decisions105106.
• Functions: Maintain internal state, deduce reasoning, update knowledge from observations, and take
actions106.
• Store knowledge about surroundings as "sentences" (technical facts)106....
4.3.1 The Architecture of Knowledge-Based Agent
• KBAs contain a knowledge-base (KB) and an inference engine (IE)109.
◦ Knowledge base (KB): Stores real-world facts using sentences in a knowledge representation
language109. The learning element regularly updates it107.
◦ Inference engine (IE): Infers new knowledge from old sentences and adds them to the KB, applying
logical rules (forward or backward chaining)108109.
• Process: KBA perceives environment, input goes to IE, IE interacts with KB for decisions107.
4.3.2 Operations Performed By KBA
• TELL operation: Informs the KB what knowledge it has or needs, and what action was
selected/performed108110.
• ASK operation: Queries the KB for what action to perform110.
• PERFORM operation: Executes the selected action110.
4.3.3 A Generic Knowledge-Based Agent
• Accepts environmental percepts as input, returns an action as output111.
• Maintains a knowledge base (KB) with background knowledge and a time counter111.
• Steps:
◦ TELLS KB what it perceives (using MAKE-PERCEPT-SENTENCE)111.
◦ ASKS KB what action to take (using MAKE-ACTION-QUERY)112.
◦ TELLS KB about the chosen action (using MAKE-ACTION-SENTENCE)112.
4.3.4 Various Levels of Knowledge-Based Agent
• Knowledge Level: Specifies what the agent knows and its goals, used to fix agent behaviour (e.g.,
optimum path from A to B)112113.
• Logical Level: Understands knowledge representation by encoding sentences into different logics (e.g.,
deducing logic for the path)113.
• Implementation Level: Physical representation of logic and knowledge; agent performs actions based
on knowledge from higher levels (e.g., actually moving from A to B)113.
4.3.5 Approaches to Designing a Knowledge-Based Agent
• Declarative Approach: Initializes with an empty KB and gradually adds sentences (facts) until it's
knowledgeable114115.
• Procedural Approach: Directly encodes desired behaviour into the agent as program code (e.g., in
LISP, Prolog)115.
• Hybrid Approach: Combines both, compiling declarative knowledge into efficient procedural code115.
4.4 Types of Knowledge Knowledge can be expressed in a KR system in different ways116:
 RV Institute of Technology and
Management®

• Simple Relational Knowledge:

◦ Facts about objects stored systematically using relations (tables) in databases116.
◦ Shortcoming: Little opportunity for inference116.
• Inheritable Knowledge:
◦ Stores data using a hierarchy of classes (generalised to specialised)116.
◦ Shows relation between instance and class (IS-A relation)116.
• Inferential Knowledge:
◦ Represents knowledge in formal logic, used to derive more facts accurately117.
◦ Example: "Diya is a student. All students are bright." -> "Student(Diya)" and "∀x = Student (x) →
Bright (x)"117