The PG Diploma course by upGrad is one of the most comprehensive ones. It covers all the knowledge of skills, concepts and tools required in the industry currently.
The syllabus is designed to make you industry ready and ace the interviews with ease.
Let’s go over the complete syllabus for in-depth detail of the coverage of our “Executive PG Programme in Machine Learning and AI”.
The course is divided into 8 main parts:
- Data Science Tool kit
- Statistics & Exploratory Data Analytics
- Machine Learning-1
- Machine Learning-2
- Natural Language Processing
- Deep Learning
- Reinforcement Learning
- Deployment and Capstone Project
Data Science Tool kit
This part is a pre-preparatory course which is essential to start the journey of Data Science and Machine Learning. The major requirements are Python, SQL and Excel as well to some extent.
This part is divided into below 6 modules:
Introduction to Python: This module covers the core Python topics assuming no prior knowledge. Understanding the structure of Python, Data Structures like lists, tuples, dictionaries, etc. is covered.
Python for Data Science: The 2 most important libraries of Python – NumPy and Pandas are covered in depth. NumPy and Pandas are essential for Data Analysis, cleaning and most of the core Data Science work.
Math for Machine Learning: Linear Algebra, Matrices, Multi-Variable Calculus and Vectors are covered in this module. These topics are a pre-requisite for understanding how ML algorithms work.
Data Visualization in Python: This module covers the dynamics of plotting graphs and trends using Python.
- Data Analysis using SQL: SQL is at the core of Data Analysis and Engineering. This module covers the basics of SQL like functions, clauses, queries and joins.
- Advanced SQL: This module covers more advanced topics like Database design, Window functions, Query Optimization, etc.
Statistics & Exploratory Data Analytics
Statistics and Data go hand in hand. Most of the Data Analysis runs statistical analysis under the hood which can then be explored further to get significant results.
This part covers below 6 modules:
- Analytics Problem Solving: This module covers the CRISP-DM framework for an overview of a Machine Learning project spanning from business understanding to deployment.
- Investment Assignment: A Data Analytics assignment as an investment banking firm employee.
- Inferential Statistics: This module covers the most important statistical concepts like Probability, Probability Distributions and the Central Limit Theorem.
- Hypothesis Testing: The what, why and hows of Hypothesis Testing are covered in this module. P-Value, different types of tests and implementation in Python.
- Exploratory Data Analysis: EDA brings out the information from the Data. This module covers Data Cleaning, Univariate/Bivariate analysis and derived metrics for ML.
- Group Project: Lending Club Case Study to find out which customers are at risk of defaulting loans.
Learn Machine learning certification from the World’s top Universities. Earn Masters, Executive PGP, or Advanced Certificate Programs to fast-track your career.
This part covers the basics of Machine Learning and some algorithms. It is essential to have a comprehensive knowledge of these before diving into more advanced topics.
It consists of 5 modules:
- Linear Regression: This module covers the basics of linear regression, its assumptions, limitations and industry applications.
- Linear Regression Assessment: A car price prediction assignment.
- Logistic Regression: Univariate and Multivariate Logistic Regression for classification ML. Implementation in Python, evaluation metrics and industry applications are covered.
- Naive Bayes: One of the easiest and most effective classification algorithms. This module covers the basics of Bayes Theorem, Naive Bayes classifier and implementation in a Spam-Ham classifier.
- Model Selection: This module covers the model selection, Bias-Variance Tradeoff, Hyperparameter Tuning and Cross-Validation which are necessary to finalize the best ML model.
This part covers more advanced topics of Machine Learning. It consists of different types of supervised and unsupervised algorithms.
The 8 modules covered are:
- Advanced Regression: This module introduces the Generalized Linear Regression and Regularized Regression techniques like Ridge and Lasso.
- Support Vector Machine (Optional): This module covers the SVM algorithm, its working, kernels and implementation.
- Tree Models: Basics of Tree models, their structure, splitting techniques, pruning and ensembles to form Random Forests are covered here.
- Model Selection-Practical Considerations: This module gives a hands-on for using model selection techniques to select the best model.
- Boosting: What are weak learners and string learners, and how can they be joined together to form a great model. Various Boosting techniques are covered here.
- Unsupervised Learning-Clustering: This module introduces Clustering, its types and implementation from scratch.
- Unsupervised Learning-Principal Component Analysis: This covers the basics of PCA, its working and implementation in Python.
- Telecom Churn Case Study: Case Study to predict Customer Churn for a telecom operator.
Natural Language Processing
Natural Language Processing(NLP) is in itself a huge field. In this NLP part, all the building blocks of text data handling are covered along with chatbots.
The 5 modules included are:
- Lexical Processing: This module covers the basics of NLP like text encoding, Regular Expressions, text processing techniques and advanced lexical techniques like Phonetic Hashing.
- Syntactic Processing: This module covers the basics of Syntactic Processing, different types of text parsing, Information Extraction and Conditional Random Fields.
- Syntactic Processing-Assignment: Implementing Syntactic processing to understand the grammatical structure of the text.
- Semantic Processing: This module introduces Semantic Processing, Word vectors and embeddings, Topic Modelling techniques followed by a case study.
- Building Chatbots with Rasa: This module covers the hottest tool for chatbot development along with implementation.
Deep Learning is widely used in the industry in many cutting edge applications for various types of data. In this part, all the types of Neural Networks are covered along with implementation.
The 5 modules covered are:
- Introduction to Neural Networks: This module covers the basics of Neural Networks, activation functions and the Feed Forward network.
- Convolutional Neural Network-Industry Applications: This module covers in detail the CNN, its structure, layers and working. It also covers various Transfer Learning models, Style Transfer and Data pre-processing of image data followed by a case study.
- Neural Networks-Assignment: A CNN based case study.
- Recurrent Neural Networks: This module covers another type of neural networks specially used for sequence-based data – RNN and LSTM along with their implementations.
- Neural Networks Project: In this module, you’ll be doing a Gesture Recognition project using CNNs and RNNs network stacks.
In this part, we introduce you to another type of Machine Learning – Reinforcement Learning. You’ll learn the basics including the classical reinforcement learning as well as Deep Reinforcement Learning.
This part covers below 4 modules:
- Classical Reinforcement Learning: This module covers the basics of RL like Markov Decision Process, RL Equations as well as Monte Carlo Methods.
- Assignment-Classical Reinforcement Learning: A tic-tac-toe assignment using RL.
- Deep Reinforcement Learning: In this module, we’ll dive into Deep Q Networks, their architecture and implementation. It also covers more advanced topics like Policy Gradient Methods and Actor-Critic Methods.
- Reinforcement Learning Project: An assignment to be done using RL architecture.
In this part, you will make your final capstone project using all the knowledge gained so far.
This part is divided into 2 modules:
- Deployment: This module covers the later stage of a Machine Learning project where you’ll learn the deployment basics on cloud and PaaS, as well as CI/CD pipelines and Docker basics.
- Capstone: The final capstone project to make your resume and portfolio skyrocket.
Before You Go
This program covers all the required basics and advanced tools and skills to enter the Data Science and Machine Learning Industry. You’ll be going through a sufficient amount of practicals and projects to make sure you’ve learnt well.
With all the learnt skills you can get active on other competitive platforms as well to test your skills and get even more hands-on.
What is machine learning?
Machine learning is a field of computer science that gives computers the ability to learn without being explicitly programmed. Giving computers the ability to learn without being explicitly programmed. Machine learning is the scientific discipline that studies the construction and study of algorithms that can learn from and make predictions on data. From the problem statement, machine learning focuses on predictive modeling from the given data/features, and forms a hypothesis about the probability of an outcome based on the features present in the data.
What are the applications of machine learning?
In general, machine learning is a kind of artificial intelligence (AI) that involves a computer or a program to learn and make predictions based on data. Machine learning is already widely used in image recognition, natural language processing and various other fields, while the recent breakthroughs in deep learning and big data have brought AI closer to reality. Currently, machine learning is being used in almost all the crucial sectors including healthcare, transport and logistics, agriculture, ecommerce, etc.
How to create a machine learning model?
A machine learning model learns from labeled training data and makes predictions or classifications on new, previously unseen data. It is based on statistical learning theory, but with a lot of optimization, modeling, and coding. A machine learning model therefore has two parts, a model and a learning algorithm. The model part is represented as a mathematical model, such as a tree or a decision-tree, and the learning algorithm is represented by a historical dataset. The learning algorithm will learn from the dataset and optimize the model to balance the error and the complexity of the model. The more accuracy your model gets and the simpler the model is, the better it is.