AIML

Dr. Shanti Chandra

Assistant Professor
Electronics and Communication Engineering
 LTP structure of the course: 3-0-0

Unit Topics for Coverage

1-Mid 1 Introduction to Machine Learning
term Basic ML concepts and examples, Basic Probability Notations, Bayesian Inference, Basic
concepts of statistics, probability and calculus.
2 Supervised Machine Learning
Regression (Linear Regression, Ridge regression, Regression Trees, Non-linear regression,
Bayesian Linear regression, polynomial regression, Lasso regression, Gradient decent )
Classification (Random forest, Decision Trees, Logistic Regression, Naïve Bayes classifier,
multi-class classification) Maximum Likelihood estimation, Regularization/MAP, Soft/Hard
Margin SVM, SVM Duality
3 Unsupervised Machine Learning
Clustering (K-means clustering (Soft/Hard), KNN (k-nearest neighbors), Hierarchical
clustering, Anomaly detection, Neural networks, Principal Component Analysis,
Independent Component Analysis, A-priori algorithm, Posteriori Algorithm, Singular value
decomposition)
Association (Hidden Markov Model, Gaussian Mixture Model, Gaussian Mixture Model-
Universal Background Model, Joint Factor Analysis, i-vector, i-vector/PLDA methods)
4 Reinforcement Machine Learning
Reinforcement Learning overview, The learning Task, Q-Learning, Nondeterministic Q-
Learning, Temporal Difference-Learning, RL-General formulation, Multi-armed Bandits,
Markov Decision Process and Deep Reinforcement Learning
Text Books and References:
1. Pattern Recognition and Machine Learning by Bishop, Springer, 2006.
2. Machine Learning: A Probabilistic Perspective by Kevin P. Murphy, MIT Press,
2012The Elements of Statistical Learning, 2nd edition by Hastie, Tibshirani and
Friedman, Springer-Verlag, 2008.
3. Bayesian Reasoning and Machine Learning by David Barber, Cambridge University
Press, 2012.
4. Information Theory, Inference, and Learning Algorithms by David Mackay, Cambridge
University Press, 2003.
5. Richard Sutton and Andrew Barto, Reinforcement Learning: An Introduction, Second
Edition, MIT Press, 2018.
What is Artificial Intelligence?

 The science and engineering of making intelligent machines, especially intelligent

computer programs
 One of the booming technologies of computer science is Artificial Intelligence which is
ready to create a new revolution in the world by making intelligent machines.

 "It is a branch of computer science by which we can create intelligent machines which
can behave like a human, think like humans, and able to make decision…………??
The Artificial Intelligence is now all around us. It is currently working with a variety of
subfields, ranging from general to specific, such as self-driving cars, playing chess, proving
theorems, playing music, Painting, etc.
Facial recognition
Digit recognition
Molecular classification
Why Artificial Intelligence?
Before Learning about Artificial Intelligence, we should know that what is the importance of AI
and why should we learn it. Following are some main reasons to learn about AI:
 With the help of AI, you can create such software or devices which can solve real-world
problems very easily and with accuracy such as health issues, marketing, traffic issues, etc.
 With the help of AI, you can create your personal virtual Assistant, such as Cortana, Google
Assistant, etc.
 With the help of AI, you can build such Robots which can work in an environment where
survival of humans can be at risk.
 AI opens a path for other new technologies, new devices, and new Opportunities.
Goals of Artificial Intelligence
Following are the main goals of Artificial Intelligence:
 Replicate human intelligence
 Solve Knowledge-intensive tasks
 An intelligent connection of perception and action
 Building a machine which can perform tasks that requires human intelligence such as:
 Proving a theorem
 Playing chess
 Plan some surgical operation
 Driving a car in traffic
 Creating some system which can exhibit intelligent behavior, learn new things by itself,
demonstrate, explain, and can advise to its user.
Advantages of Artificial Intelligence
Following are some main advantages of Artificial Intelligence:
High Accuracy with less errors: AI machines or systems are prone to less errors and high
accuracy as it takes decisions as per pre-experience or information.
High-Speed: AI systems can be of very high-speed and fast-decision making, because of that
AI systems can beat a chess champion in the Chess game.
High reliability: AI machines are highly reliable and can perform the same action multiple
times with high accuracy.
Useful for risky areas: AI machines can be helpful in situations such as defusing a bomb,
exploring the ocean floor, where to employ a human can be risky.
Digital Assistant: AI can be very useful to provide digital assistant to the users such as AI
technology is currently used by various E-commerce websites to show the products as per
customer requirement.
Useful as a public utility: AI can be very useful for public utilities such as a self-driving car
which can make our journey safer and hassle-free, facial recognition for security purpose,
Natural language processing to communicate with the human in human-language, et
Disadvantages of Artificial Intelligence
Every technology has some disadvantages, and the same goes for Artificial intelligence. Being so
advantageous technology still, it has some disadvantages which we need to keep in our mind
while creating an AI system. Following are the disadvantages of AI:
 High Cost
 Can't think out of the box
 No feelings and emotions
 Increase dependency on machines
 No Original Creativity
Difference Between Machine Learning and Artificial
Intelligence
Machine Learning
 Definition (Mitchell, 1997)
“A computer program is said to learn from experience (E) with respect to some class of
tasks (T) and performance measure (P), if its performance at tasks in T, as measured by P,
improves with experience E.”

 Experience (E): The data the algorithm sees.

 Task (T): What we want the algorithm to do
(classification, regression, etc.).
 Performance (P): accuracy, error rate, etc.

Task (T)
Classification
Classification with Missing Inputs
Regression
Transcription : Example: Speech-to-text.
Machine Translation: Example: English → German.
Structured Output: Predict outputs with structure (like graphs, sequences, maps).
Example: Detecting roads from aerial images.
Performance Measure (P)
The performance measure tells us how well a machine learning model is doing at a task, Different
tasks need different performance measures.
The Experience, (E)
Experience refers to the data the algorithm sees during learning. Different ML approaches use
data differently:

1. Unsupervised Learning
Experience (E): Dataset with only features (no labels).
The algorithm tries to discover patterns or structure.
Mathematically: learns the probability distribution p(x) from many examples of random vector x.
Examples:
Anomaly Detection: Bank transactions dataset → Detect unusual spending patterns (fraud).
Denoising: Image dataset → Learn to remove noise and reconstruct cleaner images.
Clustering: Group customers into segments based only on behavior (no pre-defined labels).
2. Supervised Learning
Experience (E): Dataset with features (x) and labels (y).
Algorithm observes pairs (x,y )and learns the relationship.
Mathematically: learns conditional probability distribution p(y∣x).
Examples:
Classification:
Input (x): Email text.
Label (y): Spam / Not Spam.
Goal: Learn p(y∣x).

3. Regression:
Input (x): House size, location, rooms.
Label (y): House price.
Goal: Predict price → Learn p(y∣x)
 Why use Machine Learning?

1. Static, pre-programmed rules give poor performance

Writing fixed rules works for simple problems (like “if it rains, carry an umbrella”).
But real-world problems are too complex and dynamic for static rules.
Example: Spam filters – spammers constantly change tricks, so fixed rules fail. ML adapts
automatically by learning from new data.

2. Difficulties in analytically specifying the system

•Some tasks are too complicated to describe with equations or rules.
•Example: Recognizing faces in images – impossible to hand-code rules for every lighting,
angle, or background. But ML can learn from examples.

3. Dimensionality of the problem is prohibitive

•Some problems have too many variables (features) for manual programming.
•Example: Predicting disease risk from genetics – thousands of genes interact in complex ways.
ML can handle such high-dimensional data.
4. Flexibility
•ML is not guaranteed to always solve a problem, but it is a powerful tool worth trying when
traditional methods fail.
•Example: Self-driving cars
• too complex to program manually,
but ML makes progress possible.

We want ML when problems are too complex, too large, or too dynamic for human-
designed rules.
It’s not magic, but it’s often the only practical way to get solutions that improve with more
data.