18CSC305J - Artificial Intelligence UNIT 4 & 5

 I have only updated some answers...The ppt is not clear and many contents are missing, so you can say we have to refer online only for the answers. So please make do with these.

4M:


Supervised vs Unsupervised


SUPERVISED LEARNING

UNSUPERVISED LEARNING

Requires labeled data

Uses unlabeled data

Predicts output based on input-output relationships

Extracts patterns or relationships from the input data

Classification and regression

Clustering and association

Receives feedback during training to adjust parameters

Does not receive explicit feedback during training

More complex

More simple


What is adaptive network?

  • Adaptive learning brings human analysts into the process at every step. This is in contrast to rule-based, simple machine learning and deep learning approaches, where the humans only create rules and label data at the start of the process. 

  • Adaptive learning systems require the least human effort because they only require human input when it matters most and continually expand their knowledge when new information is encountered. 

  • They are also the most accurate. They combine the three other types of machine intelligence, adding new types of ‘unsupervised machine learning’ and methods for optimizing the input from multiple, possibly disagreeing, humans.


Define Machine Translation

  • Machine translation is the process of using artificial intelligence to automatically translate text from one language to another without human involvement

  • Modern machine translation goes beyond simple word to word translation to communicate the full meaning of the original language text in the target language

  • It analyzes all text elements and recognizes how the words influence one another

  • Benefits: 

  • Automated translation assistance

  • Speed and volume

  • Cost effective

  • Large language selction


NLP applications

  • Chatbot

  • Voice assistants

  • Autocomplete in search engines

  • Sentiment analysis

  • Machine translation

  • Syntax and semantic analysis

  • Pragmatics and discourse analysis


12M:


Goal Stack planning

  • Push the original goal to the stack

  • Repeat until stack is empty

  • If stack top is a

  • Compound goal, push its unsatisfied subgoals to the stack

  • Single unsatisfied goal, replace with operator and push operator’s precondition to the stack

  • Operator, pop it, execute it and change the database by its effects

  • Satisfied goal, pop it

Goal Stack problem - idea:

  • Place goal in goal stack

  • Considering top as goal 1, place it’s subgoals on top of it

  • Then try to solve subgoals goal S1-2 and continue…

Goal Stack Problem - algorithm:

  • Find an operator that satisfies subgoal G1 and replace G1 with operator

  • If more than one operator satisfies G1, apply heuristics and choose one

  • To execute the top most statement, its preconditions are added onto the stack

  • Once preconditions are satisfied, then operator can be applied to a new state

  • New state is obtained by using ADD and DELETE lists of an operator

  • Problem solver keeps track of operators applied

  • This process is repeated until stack empty and problem solver returns the plan of the problem

Goal stack problem:

Initial state:

Ontable(A)

ontable(B)

on(D,A)

on(C,D)

clear(C)

clear(B)

Handempty 


Goal state:

Ontable(B)

ontable(C)

on(D,B)

on(A,D)

clear(C)

clear(B)

Handempty 


Complete solution:


Mycin, Dendral, XCON

Mycin:

  • MYCIN was an early expert system that used artificial intelligence to identify bacteria causing severe infections

  • It was also used for the diagnosis of blood clotting diseases

  • Also recommended antibiotics, with dosage adjusted based on patient’s weight

  • Developed at the Stanford University in the early 1970s

  • It was written in Lisp

  • It was a standalone system that required the user to enter all relevant information about the patient that MYCIN asks

  • It would ask the doctor some simple yes/no or textual questions

  • Used a simple inference engine and a knowledge base of ~600 rules


XCON:

  • The R1 was a production rule based expert system written in OPS5 by John McDermott in 1978

  • It performed ordering of DEC’s VAX computer systems by automatically selecting the computer system components based on the customer’s requirements

  • It helped by reducing the need to give customers free components when technicians made errors, by speeding the assembly process and by increasing customer satisfaction

  • It had about 2500 rules

  • XCON’s success made DEC rewrite XCON as XSEL for use by DEC’s Salesforce


DENDRAL:

  • Dendral was an influential pioneer expert system

  • The name is a short form of ‘Dendritic algorithm’

  • It was written in Lisp

  • Developed at Stanford University

  • The main aim was to help organic chemists in identifying unknown organic molecules, by analyzing their mass spectra and using knowledge of chemistry

  • Parts of a dendral:

  • Heuristic dendral (graph theory algorithm)

  • Meta dendral (machine learning)


CNN:

  • A Convolutional Neural Network is a type of deep learning neural network architecture commonly used in computer vision

  • CNN consists of multiple layers like the input layer, convolutional layer and fully connected layer

  • CNN is mainly used for

  • Image recognition

  • Image classification

  • Object detection

Types of layers in CNN:

Input layer:

  • The layer in which we give input to our model

  • Generally the input will be an image or a sequence of images


Convolution layer:

  • An image matrix of dimension  h x w x d

  • A filter fh x fw x fd

  • Outputs a volume dimension

  • Layer which is used to extract the feature from the input dataset

  • It applies a set of filters to the input images


Example: applying filters

The result matrix is 

6 -9 -8

-3 -2 -3

-3 0 -2


Activation layer:

  • Activation layers add nonlinearity to the network

  • Apply an element wise activation function to the output of the convolution layer


Pooling layer:

  • Layer is periodically inserted in the CNN and main function is to reduce the size of volume 

  • This makes computation fast, reduces memory and prevents overfitting

  • Two common types:

  • Max pooling

  • Average pooling


Advantages:

  • Good at detecting patterns and features in images, videos and audio signals

  • No need for manual feature extraction

  • Can handle large amount of data and give high accuracy

  • Robust to translation, rotation and scaling invariance

Disadvantages: 

  • Computationally expensive

  • Requires large amount of labeled data

  • Can be prone to overfitting

  • Interpretability is limited



Comments

Popular posts from this blog

18CSC303J - Database Management System UNIT 4 & 5 - 12 MARKS

18CSC303J - Database Management System UNIT 4 & 5 - 4 MARKS