This block of Python for AIML online course will teach you all about the Exploratory Data Analysis like Preprocessing, Missing values, etc.

• • Data, Data Types, and Variables
    This module will drive you through some essential data types and variables.
  • Central Tendency and Dispersion
    Central tendency is expressed by median and mode. Dispersion is described by data that is distributed around this central tendency. Dispersion is represented by a range, deviation, variance, standard deviation and standard error.
  • 5 point summary and skewness of data
    5 point summary is a large set of descriptive statistics, which provides information about a dataset. Skewness characterises the degree of asymmetry of a distribution around its mean.
  • Box-plot, covariance, and Coeff of Correlation
    This module will teach you how to solve the problems of Box-plot, Covariance, and Coefficient of Correlation using Python.
  • Univariate and Multivariate Analysis
    Univariate Analysis and Multivariate Analysis are used for statistical comparisons.
  • Encoding Categorical Data
    You will learn how to encode and transform categorical data using Python in this module.
  • Scaling and Normalization
    In Scaling, you change the range of your data. In normalisation, you change the shape of the distribution of your data.
  • What is Preprocessing?
    The process of cleaning raw data for it to be used for machine learning activities is known as data pre-processing. It’s the first and foremost step while doing a machine learning project. It’s the phase that is generally most time-taking as well. In this module, you will learn why is preprocessing required and all the steps involved in it.
  • Imputing missing values
    Missing values results in causing problems for machine learning algorithms. The process of identifying missing values and replacing them with a numerical value is known as Data Imputation.
  • Working with Outliers
    An object deviating notably from the rest of the objects, is known as an Outlier. A measurement or execution error causes an Outlier. This module will teach you how to work with Outliers.
  • “pandas-profiling” Library
    The pandas-profiling library generates a complete report for a dataset, which includes data type information, descriptive statistics, correlations, etc.

This block will teach you how to predict future values based on the previously experimented values using Python

• • Introduction to forecasting data
    In this module, you will learn how to collect data and predict the future value of data focusing on its unique trends. This technique is known as Forecasting data.
  • Definition and properties of Time Series data
    This module will teach you about the introduction of time series data and cover all the time-series properties.
  • Examples of Time Series data
    You will learn some real-time examples of time series data in this module.
  • Features of Time Series data
    You will learn some essential features of time series data in this module.
  • Essentials for Forecasting
    In this module, you will go through all the essentials required to perform Forecasting of your data.
  • Missing data and Exploratory analysis
    Exploratory Data Analysis, or EDA, is essentially a type of storytelling for statisticians. It allows us to uncover patterns and insights, often with visual methods, within data. In this module, you will learn the basics of EDA with an example.
  • Components of Time Series data
    In this module, you will go through all the components required for Time-series data.
  • Naive, Average and Moving Average Forecasting
    Naive Forecasting is the most basic technique to forecast your data like stock prices. Whereas, Moving Average Forecasting is a technique to predict future value based on past values.
  • Decomposition of Time Series into Trend, Seasonality and Residual
    This module will teach you how to decompose the time series data into Trend, Seasonality and Residual.
  • Validation set and Performance Measures for a Time Series model
    In this module, you will learn how to evaluate your machine learning models on time series data by measuring their performance and validation them.
  • Exponential Smoothing method
    A time series forecasting method used for univariate data is known as the Exponential Smoothing method, one of the most efficient forecasting methods.
  • ARIMA Approach
    ARIMA stands for Auto Regression Integrated Moving Average and is used to forecast time series following a seasonal pattern and a trend. It has three key aspects, namely: Auto Regression (AR), Integration (I), and Moving Average (MA).

This block will teach you all the prerequisites you need to know before learning Deep Learning.

• • Mathematics for Deep Learning (Linear Algebra)
    This module will drive you through all the essential concepts like Linear Algebra required for implementing Mathematicsto implement mathematics.
  • Functions and Convex optimization
    Convex Optimization is like the heart of most of the ML algorithms. It prefers studying the problem of minimising convex functions to convex sets. This module will teach you how to use all the functions and convex optimisation for your ML algorithms.
  • Loss Function
    Loss Function is a technique for prediction error of neural networks.
  • Introduction to Neural Networks and Deep Learning
    This module will teach you everything you need to know about the introduction to Neural Networks and Deep Learning.

This block will teach you how to deploy your machine learning models using Docker, Kubernetes, etc.

• • Model Serialization
    Serialization is a technique to convert data structures or object state into a format like JSON, XML, which can later be stored or transmitted and reconstructed.
  • Updatable Classifiers
    This module will teach you how to use updatable classifiers for machine learning models.
  • Batch mode
    Batch mode is a network roundtrip-reduction feature. It is used to batch up data-related operations to perform them in coarse-grained chunks.
  • Real-time Productionalization (Flask)
    In this module, you will learn how to improve your Machine Learning model’s productivity Using Flask.
  • Docker Containerization – Developmental environment
    Docker is one of the most popular tools to create, deploy, and run applications with the help of containers. Using containers, you can package up an application with all the necessary parts like libraries and other dependencies, and ship it all together as one package.
  • Docker Containerization – Productionalization
    In this module, you will learn how to improve the productivity of deploying your Machine Learning models.
  • Kubernetes
    Kubernetes is a tool similar to Docker that is used to manage and handle containers. In this module, you will learn how to deploy your models using Kubernetes.

This block will teach you how TensorBoard provides the visualization and tooling required for machine learning experimentation.

• • Callbacks
    Callbacks are powerful tools that help customise a Keras model’s behaviour during training, evaluation, or inference.
  • Tensorboard
    TensorBoard is a free and open-source tool that provides measurements and visualizations required during the machine learning workflow. This module will teach you how to use the TensorBoard library using Python for Machine Learning.
  • Graph Visualization and Visualizing weights, bias & gradients
    In this module, you will learn everything you need to know about Graph Visualization and Visualizing weights, bias & gradients.
  • Hyperparameter tuning
    This module will drive you through all the concepts involved in hyperparameter tuning, an automated model enhancer provided by AI training.
  • Occlusion experiment
    Occlusion experiment is a method to determine which image patches contribute to the maximum level to the output of a neural network.
  • Saliency maps
    A saliency map is an image, which displays each pixel’s unique quality. This module will cover how to use a saliency map in deep learning.
  • Neural style transfer
    Neural style transfer is an optimization technique that takes two images, a content image and a style reference image, later blends them together. Now, the output image resembles the content image but displayed in the style of the style reference image.

This block will teach you how to implement GANs (Generative Adversarial Networks) in Machine Learning.

• • Introduction to GANs
    Generative adversarial networks, also known as GANs, are deep generative models. Like most generative models they use a differential function represented by a neural network known as a Generator network. GANs also consist of another neural network called Discriminator network. This module covers everything about the introduction to GANs.
  • Autoencoders
    Autoencoder is a type of neural network where the output layer has the same dimensionality as the input layer. In simpler words, the number of output units in the output layer is equal to the number of input units in the input layer. An autoencoder replicates the input to the output in an unsupervised manner and is sometimes referred to as a replicator neural network.
  • Deep Convolutional GANs
    Deep Convolutional GANs works as both Generator and Discriminator. You will learn how to use Deep Convolutional GANs with an example.
  • How to train and common challenges in GANs
    In this module, you will learn how to train GANs and identify common challenges in GANs.
  • Semi-supervised GANs
    The Semi-Supervised GAN is used to address semi-supervised learning problems.
  • Practical Application of GANs
    In this module, you will learn all the essential and practical applications of GANs.

This block will cover all the essential aspects of Reinforcement Learning used in various Machine Learning applications.

• • What is reinforcement learning?
    We need technical assistance to simplify life, improve productivity and to make better business decisions. To achieve this goal, we need intelligent machines. While it is easy to write programs for simple tasks, we need a way to build machines that carry out complex tasks. To achieve this is to create machines that are capable of learning things by themselves. Reinforcement learning does this.
  • Reinforcement learning framework
    You will learn some essential frameworks used for Reinforcement learning in this module.
  • Value-based methods – Q-learning
    The ‘Q’ in Q-learning stands for quality. It is an off-policy reinforcement learning algorithm, which always tries to identify the best action to take provided the current state.
  • Exploration vs Exploitation
    Here, you will discover all the key differences between Exploration and Exploitation used in Reinforcement learning.
  • SARSA
    SARSA stands for State-Action-Reward-State-Action. It is an on-policy reinforcement learning algorithm, which always tries to identify the best action to take from another state.
  • Q Learning vs SARSA
    Here, you will discover all the key differences between Q Learning and SARSA used in Reinforcement learning.

The recommendation system is the application used to study the user’s former actions and recommend the most accurate recommendations with regards to their historical behaviour and preferences. These recommendations intensify the user’s experience. We learn what recommendation systems are, their applications, key approaches to building them – Popularity based systems, Collaborative filtering, Singular Value Decomposition etc. Throughout this module, candidates would participate in 3 Quizzes and work on 1 Project to enhance their understanding.
Below are the various concepts of Recommendation systems you would master

• • Introduction to Recommendation systems: You would gain a basic understanding of Recommendation Systems.                              
  • Content based recommendation system: You would learn the techniques to filter the content and recommend the best products.
  • Popularity based model: You would learn the techniques of popularity-based filtering of the recommendations.
  • Collaborative filtering (User similarity & Item similarity): collaborative filtering is a process of making automatic predictions of the interests of a user by collecting preferences or taste information from several users. You would learn the techniques of collaborative-based filtering of the recommendations.
  • Hybrid models Hybrid systems blend neural networks and Machine Learning techniques to recognize the patterns in a given dataset. They are rigorously employed in solving the problems of deep learning.