iis - Module 6 Hard capped

What is an Expert System?

An AI program that imitates the decision-making of a human expert in a narrow domain using a knowledge base and inference engine.

Give examples of Expert Systems from the notes.

MYCIN → medical diagnosis; DENDRAL → chemical analysis; XCON → computer configuration; PROSPECTOR → mineral exploration.

Why are Expert Systems needed?

They preserve expert knowledge, provide fast and consistent decisions, can operate where human expertise is scarce, expensive or unavailable, can work in hazardous environments, and may explain their reasoning.

What are the main components of an Expert System?

Knowledge Base, Inference Engine, Working Memory, User Interface, Explanation Facility, and Knowledge Acquisition.

What is the role of the Knowledge Base in an Expert System?

It stores domain knowledge, including facts and IF-THEN rules.

What is the role of the Inference Engine?

It applies rules to facts and derives conclusions.

What is Working Memory in an Expert System?

It stores the current facts or information relevant to the problem being solved.

What is the basic Expert System architecture?

User ↔ User Interface ↔ Inference Engine; the Inference Engine interacts with the Knowledge Base and Working Memory, while the Explanation Facility explains reasoning and Knowledge Acquisition supplies domain knowledge.

What is the role of the Explanation Facility and Knowledge Acquisition component?

The Explanation Facility provides reasoning/explanation to the user; Knowledge Acquisition obtains and incorporates domain knowledge into the system.

What is Machine Learning (ML)?

Machine Learning is an approach in which systems learn patterns from data rather than being explicitly programmed for every decision.

What are the main learning types relevant to Module 6?

Supervised learning uses labelled data, unsupervised learning discovers patterns in unlabelled data, and reinforcement learning learns through rewards and penalties.

What is Deep Learning (DL)?

Deep Learning is machine learning based on neural networks with many hidden layers.

What is the relationship between AI, ML and DL?

AI is the broader field → ML is a subset of AI → DL is a subset of ML based on deeper neural networks.

What are the major Deep Learning architectures listed in the notes?

CNN, RNN/LSTM, and Transformer.

How does a neural network learn according to the notes?

Input → forward propagation → output → calculate error → backpropagation → gradient descent → update weights → repeat over many epochs.

What are the key points to write for a 5-mark Deep Learning answer?

Definition → basic architecture → how it works → backpropagation and gradient descent → architectures → advantages → disadvantages → applications.

What is an Artificial Neural Network (ANN)?

An ANN is a network of weighted artificial neurons that processes inputs using weights, bias and activation functions.

What is the basic artificial neuron formula?

net = Σwᵢxᵢ + b and y = f(net), where weighted inputs plus bias are passed through an activation function.

What is the purpose of an activation function in an ANN?

The activation function determines the neuron's output from its weighted input.

What is the complete basic ANN learning flow?

Input → Forward propagation → Output → Calculate error → Backpropagation → Gradient descent → Update weights → Repeat.

What is Backpropagation?

A learning process that works backward from the error to determine how the network's weights contributed to that error so the weights can be adjusted.

What is Gradient Descent in ANN learning?

A method used to adjust the network's weights in a direction that reduces the error.

What should be included in a 10-mark ANN answer according to the notes?

Definition → ANN architecture diagram → artificial neuron → net = Σwᵢxᵢ+b → y=f(net) → activation functions → forward propagation → error → backpropagation → gradient descent → types → advantages → disadvantages → applications.

What is Soft Computing?

A broader family of techniques designed to handle uncertainty or imprecision; the notes place fuzzy logic and ANN within this broader family.

What is Fuzzy Logic?

A logic approach in which truth can take values between 0 and 1 rather than being restricted to only completely true or completely false.

What is the basic Fuzzy System pipeline?

Fuzzification → Fuzzy Rules → Inference → Defuzzification → Crisp Output.

Explain the washing-machine example of a Fuzzy System.

The input “clothes dirty” is fuzzified; for example, “very dirty” may have membership 0.8. A rule such as IF clothes are very dirty THEN wash for long time is applied through inference, then defuzzification produces a crisp output such as a 45-minute wash time.

What are applications of Fuzzy Logic given in the notes?

Washing machines, air conditioners, and camera autofocus.

What is Natural Language Processing (NLP)?

NLP deals with understanding and generating human language.

What are the five phases of the NLP pipeline?

Lexical analysis → Syntactic analysis → Semantic analysis → Discourse integration → Pragmatic analysis.

What are Robotics and Generative AI in the context of the notes?

Robotics follows the basic Sense → Plan → Act cycle. Generative AI refers to AI models that learn patterns from data to generate new content such as text, images, audio and code.