Professional Documents
Culture Documents
AML - Lab - Syllabus - Chandigarh University
AML - Lab - Syllabus - Chandigarh University
Course Description
The objective of this course is to learn and understand advanced topics in the field of machine learning and its
applications. This course gives a graduate-level advancement in machine learning and in-depth coverage of new
and advanced algorithms in machine learning, as well as their underlying theory. The course assumes basic
knowledge of machine learning, and a good background in probability and statistics, linear algebra, and
optimization methods. It emphasizes approaches with practical relevance and discusses a number of recent
applications of machine learning in areas like Deep Learning, Information retrieval, recommender systems, data
mining, computer vision, natural language processing and robotics. This course aims to delve deeper into the
intricacies of various algorithms and methodologies used in advanced machine learning applications. This
course will teach students how to investigate data from beginning to end using a machine learning perspective.
It will teach students a few of the most crucial machine learning algorithms, how to extract and select valuable
features that best reflect your data, and how to assess the effectiveness of your machine learning algorithms.
COURSE OBJECTIVES
• To provide a comprehensive understanding of advanced machine learning algorithms beyond the basics.
• To develop skills in using recent machine learning software for solving practical problems
• To expose students to the latest developments and applications of advanced machine learning in various
fields like computer vision, natural language processing, speech recognition, recommendation systems,
and autonomous agents
• To provide students with a solid theoretical understanding of the principles behind advanced Machine
learning algorithms.
• To Foster critical thinking and research skills in students and to apply machine learning to real-world
problems
Course Outcomes
On completion of this course, the students shall be able to
Identify/Recognize the feasibility of applying a of advanced machine learning algorithms for a
CO1
particular problem
Identify cutting-edge applications of machine learning and effectively use modern software tools
CO2
to solve real problems based on machine learning.
Compare and analyze different advanced Machine learning Models based on evaluation
CO5
parameters.
Advanced Machine Learning Lab on the basis of modules
This lab will be carried out in three modules and five graded and four non graded experiments. Every
module contains a real-life problem that can be solved by machine learning algorithms. Students will
download datasets from standard repositories and will build a machine-learning model. Students will
work from applying feature engineering to UX/UI designing. Evaluation will be done on the basis of
the given rubrics.
List of modules:
Unit-1 Contact Hours: 4 hrs
Experiment 6 Implement a basic CNN architecture with convolutional and pooling layers for image
(Non-Graded) classification on the CIFAR-10 dataset.
Students will download dataset of Handwritten English Alphabets from the repository
https://www.kaggle.com/datasets/sachinpatel21/az-handwritten-alphabets-in-csv-format
and will perform following task.
Experiment 7
1. Write a script to convert .CSV file to actual images in .PNG format in structured
(Graded)
folder.
2. Design a handwritten English alphabet recognition system using CNN
Students will download dataset from Kaggle repository Choose a dataset that consists of a
large text corpus. Examples include books, articles, or even song lyrics. Preprocess the
dataset by cleaning the text, removing unnecessary characters or symbols, and splitting it
into training and validation sets. Implement an RNN model architecture using either
LSTM (Long Short-Term Memory Train the RNN model using the training dataset,
Experiment 8
adjusting hyperparameters such as learning rate, batch size, and number of training
(Graded)
epochs.
1. Evaluate the performance of the trained model using the validation dataset.
2. Calculate metrics to assess the quality of generated text.
Unit-3 Contact Hours:4 hrs
Experiment 9
Case study on attention model for the computer vision task.
(Non-Graded)
Students need to work on how to apply advanced applications of Deep Learning
techniques in real world. Student would be encouraged to know more about Computer
Vision Libraries and how they are implemented in Python. Students need to work on
cancer data and design a prediction model on it (dataset can be downloaded from standard
Experiment 10 repositories (Kaggle, UCI machine learning, Github).
(Graded)
1. Detect an Object from cancer data (cancer using OpenCV or imagenet)
2. Perform the classification of detected object with performance parameters.
TEXT BOOKS
T1: Mitchell T.M., Machine Learning, McGraw Hill (1997).
T2: Andreas C. Miller, Sarah Guido, Introduction to Machine Learning with Python, O’REILLY (2001).
T3: Deep Learning" by Ian Goodfellow, Yoshua Bengio, and Aaron Courville
T4:Deep Learning with Python François Chollet 1 Manning Publications 2017
REFERENCE BOOKS
R1: Bishop C., Pattern Recognition and Machine Learning, Springer-Verlag (2006).
R2: Deep Reinforcement Learning Pieter Abbeel and John Schulman MIT Press 2020
CO-PO MAPPING:
Course
PO PO PO PO PO PO PO PO PO PO1 PO1 PO1 PSO PSO PSO
Outcom
1 2 3 4 5 6 7 8 9 0 1 2 1 2 3
e
CO1 - 3 - 1 - - - - 2 - 1 2 1 2 3
CO2 - 3 - - 3 - 1 3 2 - 1 2 2 2 3
CO3 3 - 3 - 2 2 1 1 2 - 1 3 3 2 2
CO4 - 2 2 - 2 1 1 - 2 - 1 3 2 3 2
CO5 - - 2 3 2 1 1 - 2 3 1 3 3 3 3
The problem
The problem statement is
The problem The problem
statement is challenging, involves The problem statement is highly
statement is statement is well-
Problem Selection and defined, but it may complex data, and innovative, addressing cutting-edge
unclear, lacks defined and
Definition 5 lack complexity or demonstrates an challenges in the field of machine
relevance, or is too relevant to
real-world understanding of learning.
simplistic. advanced machine
relevance. advanced ML .
learning concepts
techniques.
Exhibits strong
Struggles with Demonstrates some Applies problem-
problem-solving
problem-solving and problem-solving solving techniques Demonstrates outstanding problem-solving
Problem Solving and abilities, using
5 lacks creativity in skills but mostly effectively and shows and creativity, proposing novel solutions and
Creativity innovative approaches
approaching follows conventional some creativity in exploring new avenues
to tackle complex
challenges. approaches certain scenarios
challenges.
Conceptual Understanding Demonstrates a limited understanding of Shows a solid understanding of advanced Displays a deep and comprehensive understanding of
advanced ML principles and struggles to machine learning principles and can advanced ML principles and can articulate complex
explain key concepts clearly. explain most concepts with clarity concepts clearly and concisely
Theoretical Knowledge Struggles to apply underlying Demonstrates a satisfactory application of Applies advanced mathematical and statistical
mathematics and statistics to advanced mathematical and statistical principles to concepts proficiently to derive and analyze ML
ML algorithms effectively. understand advanced ML algorithms algorithms accurately.
Algorithm Implementation Struggles to implement advanced ML Successfully implements most advanced Implements advanced ML algorithms proficiently,
algorithms and produces code with ML algorithms with correct and functional producing efficient, well-organized, and optimized
significant errors or inefficiencies. code. code.
Design and Analysis Struggles to design proper experiments Designs and conducts appropriate Demonstrates exceptional experimental design skills,
and analyze results effectively. experiments, and analyzes results and conducts in-depth analysis with insightful
reasonably well. conclusions.
.
Feature Engineering and Struggles to perform feature engineering Applies standard feature engineering Demonstrates creativity and domain expertise in
Selection and selection, leading to suboptimal techniques and performs adequate feature feature engineering and selection, enhancing model
model performance. selection performance significantly.
Model Training and Tuning Faces challenges in training models and Successfully trains models with Expertly trains models, using advanced
lacks effective hyperparameter tuning appropriate tuning, achieving reasonable hyperparameter tuning techniques to achieve state-of-
skills. performance. the-art performance.
Handling Complex Data Struggles with complex data types and Adequately handles complex data types Shows exceptional skills in working with complex
Types lacks proficiency in implementing using appropriate algorithms and data, implementing advanced algorithms for specific
relevant algorithms. preprocessing techniques data types.
Transfer Learning and Demonstrates limited understanding and Successfully applies transfer learning Expertly adapts and extends transfer learning
Domain Adaptation application of transfer learning concepts. techniques in various scenarios. approaches, achieving outstanding results in
challenging domains.