Artificial Intelligence

1) Compare and contrast the following: Forward Vs. Backward

Reasoning.

2) Explain Bayesian Network.
 Bayesian Network (BN)

A Bayesian Network is a graphical model that represents variables and their conditional
dependencies using a Directed Acyclic Graph (DAG). It allows probabilistic inference and is
widely used in fields like AI, machine learning, and statistics.
Key Components:
1. Nodes: Represent random variables (discrete or continuous).
2. Edges: Directed arrows between nodes indicate conditional dependencies.
3. Conditional Probability Tables (CPTs): Quantify the effect of parent nodes on the current
node.
4. DAG (Directed Acyclic Graph): Shows relationships without cycles, ensuring clear
information flow.
Example:
Variables:
- A: Weather (Sunny, Rainy)
- B: Traffic (Heavy, Light)
- C: Accident (Yes, No)

Structure: Weather → Traffic; both influence Accident.

Conditional Probabilities (sample):

- P(A=Sunny) = 0.7, P(A=Rainy) = 0.3
- P(B=Heavy | A=Sunny) = 0.2, P(C=Yes | A=Sunny, B=Heavy) = 0.05
Properties:
1. Graphical Representation: Easier visualization of dependencies.
2. Compact Representation: Avoids the need for enumerating all probability combinations.
3. Modularity: Easy to add variables or update probabilities without altering the whole
network.
4. Inference: Efficient posterior computation using techniques like belief propagation or
MCMC.
Applications:
1. Medical Diagnosis: Linking symptoms with diseases.
2. Risk Assessment: Evaluating risks in finance/insurance.
3. NLP: Modeling word relationships for tasks like translation.
4. Machine Learning: Used in classification, clustering.
5. Robotics: Decision-making under uncertainty.

3) What is fuzzy set? Explain degree of membership. How does are represent a fuzzy set
in a computer.
 Definition: Fuzzy sets extend classical set theory by allowing varying degrees of
membership, with values ranging from 0 (not belonging) to 1 (fully belonging). This approach
helps manage uncertainty and ambiguity in real-world problems.
Key Features:
1. Partial Membership: Elements can belong to a set to varying degrees, rather than just being
in or out.
2. Membership Function: A function assigns a degree of membership (0 to 1) to each element,
indicating its strength of belonging.
3. Non-Binary Representation: Fuzzy sets are ideal for ambiguous categories (e.g., "tall,"
"hot"), capturing real-world nuances.
4. Flexibility: Supports approximate reasoning, allowing terms like "somewhat tall" to be
modeled effectively.

Degree of Membership:
- Values:
- 0: Not in the set.
- 1: Fully in the set.
- Between 0 and 1: Partial membership (e.g., height categorization).
- Example:
- "Tall People" Fuzzy Set:
- Person A (170 cm): 0.4 (somewhat tall)
- Person B (180 cm): 0.7 (tall)
- Person C (190 cm): 1.0 (very tall)
Computer Representation:
1. Fuzzy Set Definition: Stored in programming languages, with elements and their
membership values.
2. Membership Function: Calculates degrees of membership using rules or formulas.
3. Data Structures: Implemented using arrays, lists, or dictionaries.
4. Operations:
- Union: Max membership value.
- Intersection: Min membership value.
- Complement: 1 - membership value.
5. Fuzzy Logic Inference: Supports rules like “IF temperature is high THEN fan speed is fast.”
Advantages:
1. Handling Uncertainty: Effective in fields like control systems and decision-making.
2. Modeling Real-Life Scenarios: Ideal for phenomena without strict boundaries (e.g.,
temperature control).
3. Versatility: Applicable in AI, machine learning, medical diagnosis, and risk assessment.

4) What is Natural Language Processing? Explain various types of NLP

Techniques.
 Natural Language Processing (NLP):
NLP is a subfield of artificial intelligence and linguistics focused on the interaction between
computers and humans through natural language. Its goal is to enable computers to understand,
interpret, and generate human language meaningfully, aiding in text analysis, sentiment
analysis, language translation, and more.
Types of NLP Techniques:
1. Tokenization:
o Definition: Breaking down text into smaller units (tokens).
o Purpose: Segments text for easier analysis.
o Example: "NLP is amazing!" → ["NLP", "is", "amazing", "!"]
2. Part-of-Speech Tagging (POS Tagging):
o Definition: Assigning grammatical categories to tokens.
o Purpose: Understands the syntactic structure of sentences.
 o Example: "The cat sits on the mat." → "The" (determiner), "cat" (noun), "sits"
(verb).
3. Named Entity Recognition (NER):
o Definition: Identifying named entities in text (people, organizations, locations,
etc.).
o Purpose: Extracts meaningful information.
o Example: "Barack Obama was the president of the United States." → "Barack
Obama" (person), "United States" (location).
4. Sentiment Analysis:
o Definition: Determining the emotional tone of text.
o Purpose: Understands opinions and sentiments.
o Example: "I love this product!" → Positive sentiment.
5. Text Classification:
o Definition: Assigning categories to text documents.
o Purpose: Used for spam detection, topic categorization.
o Example: Classifying an email as "spam" or "not spam."
6. Machine Translation:
o Definition: Automatically translating text between languages.
o Purpose: Facilitates cross-language communication.
o Example: "Bonjour" (French) → "Hello" (English).
7. Language Modeling:
o Definition: Predicting the next word or sequence based on context.
o Purpose: Essential for text generation and speech recognition.
o Example: "The cat is on the" → Predicts "roof."
8. Text Summarization:
o Definition: Producing a concise summary of a larger text.
o Purpose: Helps digest large information quickly.
o Example: Summarizing a news article.
 9. Coreference Resolution:
o Definition: Identifying expressions that refer to the same entity.
o Purpose: Improves context understanding.
o Example: "John went to the store. He bought some milk." → "He" refers to
"John."
10. Dependency Parsing:
o Definition: Analyzing grammatical structure and relationships between words.
o Purpose: Understands syntactic structure and meaning.
o Example: "The dog chased the ball." → Subject ("dog"), verb ("chased"), object
("ball").
11. Text Generation:
o Definition: Automatically creating coherent text based on prompts.
o Purpose: Used in chatbots and content generation.
o Example: Generating a continuation of a given text.

5) What is Expert system? Explain its Architecture, Features & applications.

 An expert system is a type of artificial intelligence that mimics the decision-making ability
of human experts in specific domains. It utilizes knowledge and inference procedures to solve
complex problems through logical reasoning, primarily represented as if-then rules.
Applications include medical diagnosis, financial forecasting, and technical troubleshooting.
Key Components:
1. Knowledge Base: Domain-specific facts and rules (e.g., medical symptoms).
2. Inference Engine: Applies rules to deduce new information (e.g., evaluates symptoms).
3. User Interface: Allows user interaction for input/output (e.g., GUI for finance).
4. Explanation Facility: Justifies conclusions to enhance trust (e.g., explains diagnoses).
5. Knowledge Acquisition Module: Updates knowledge base with new information (e.g.,
medical guidelines).
Features:
 Domain Specificity: Focused on specific tasks.
 High Performance: Performs complex tasks faster than humans.
 Reliability: Provides accurate solutions.
 Explanatory Ability: Explains reasoning for conclusions.
 User-Friendly: Intuitive interfaces for easy use.
 Adaptability: Updates with new knowledge.
Applications:
1. Healthcare: Diagnosing diseases (e.g., MYCIN).
2. Finance: Investment analysis and forecasting.
3. Manufacturing: Process control and fault diagnosis.
4. Customer Support: Troubleshooting technical issues.
5. Agriculture: Pest management and crop advisory.
6. Education: Personalized learning systems.
7. Legal Systems: Legal reasoning and advice.
6) Comparison between Syntactic Processing & Semantic Processing.
