This course is specially designed to teach data science to everyone in 90 classes with no prerequisites at all so that a person with any background whether it be finance, biology, engineering can learn data science due to it’s multi domain application
Ideal student for the course
Mental Requirements
Somebody who is willing to be dedicated during the course duration, can watch videos and grasp the knowledge as the content is made super easy for everyone. Someone who can watch videos and do assignments without delay with discipline.
Physical Requirements
Someone who is looking to build a career in Data Analyst, Business Analyst, Machine Learning Engineer, Data Scientist, Deep Learning Engineer, Computer Vision, NLP , Robotics, Artificial Intelligence, Data Analysis can join this course as we cover from the very basics to very advance so no prerequisites.
All you need is a goal of what you want to achieve after this course and dedication if both meets, enroll yourself
Unique Value Proposition
No prerequisites
Everything is discussed from scratch
Even python programming is included in course
No Mathematics background is required
We’ve taught students from biology, finance, non-engineering background also, students who have never used laptop before also learned from us
Course Curriculum
- Intro to AI, ML, DS and all the course curriculum, future aspects
- Intro to Probability, Statistics, Application of Programming in ML/DS/AI
- Explanation of syntax and things in Python
- Introduction to Prediction
- Explanation of Linear Regression word
- Explanation of Line Equation
- Explanation of how prediction will be done using a line
- Explanation of Cost Function
- Calculus Module
- Intro to calculus
- Intro to Differential Calculus and applications
- Basic rules in Calculus
- Chain Rule
- Gradients and Partial Derivatives
- Linear Regression(Continued)
- Explanation of Gradient Descent
- Minimizing the cost function using Gradient Descent
- Coding Linear Regression in Python
- Quick introduction to Matrices, Vectors, Dot Product
- Multivariate Linear Regression
- Explanation of why we need Multivariate Linear Regression
- Explanation of hypothesis function
- Making hypothesis function using a dot product in Python
- Cost function and its derivatives
- Again quick explanation of Gradient Descent
- Explanation of making Gradient Function with matrices in Python
- Coding Multivariate Linear Regression in Python from Scratch
- Polynomial Regression
- Explaining the need of Polynomial Regression
- Explaining the Hypothesis Function
- Making hypothesis function using a dot product in Python
- Cost function and its derivatives
- Gradient descent in Polynomial Regression
- Coding Polynomial Regression in Python from Scratch
- Overfitting and Regularisation
- Explain the problem of overfitting
- Explain the solution with Regularisation
- Explain Variance and Standard Deviation
- Relate variance with cost function and explain what happens to variance of distribution of target variable when cost increases or decreases
- Naive Bayes Algorithm
- Numbers to Data
- Data to relative frequency
- Relative frequency to PDF
- PDF estimation through ML and achieving five qualities of AI
- Indirect vs direct methods (ML vs Probabilistic)
- Naive Bayes as direct method for estimation of PDF and classification
- Explain conditional probability
- Prove Theorem of total probability, Multiplication theorem of probability
- Prove Bayes theorem
- Explain how Bayes theorem can be used for classification
- Explain how to compute all the types of probabilities in Bayes formula
- Maximum likelihood estimation for distribution estimation
- Code Naive Bayes Classifier(Univariate and Multivariate)
- Logistic Regression
- Explain log odds
- Derive Sigmoid function from Bayes theorem/Conditional Probability
- Explain/Derive the logistic regression cost function(Cross Entropy/MLE)
- Code Logistic Regression in Python
- Linear Algebra
- Vectors and matrices explanation
- Unit vectors, Basis Vectors
- Vector Space
- Determinant of a matrix
- Inverse of a matrix
- Column Space and Null Space
- Dot Product, Cross Product
- Linear Transformations
- Singular matrix
- Dependent and Independent vectors
- EigenVectors and EigenValues
- EigenValue Decomposition(EVD)
- Singular Value Decomposition(SVD)
- Principal Component Analysis
- Issue with dependent features in data(Explain via concept of covariance)
- Recall EVD, Projection of vectors
- Explain how to project to basis vectors
- Explain Projection to rotated axes
- Explain how to make Covariance Matrix Sparse
- Explain with theorems the PCA and minimizing squares sums
- Code PCA(From Scratch)
- Support Vector Machines
- Introduction to Hyperplanes
- Intro to SVM problem via road example
- Formulate SVM Primal problem
- Formulate Dual problem
- Lagrange Function
- Dual problem using Lagrange function
- KKT Conditions
- Coding SVM using Scikit Learn
- Introduction to Deep learning:
- What is a Neural Network?
- Understanding the math behind neural network
- Weights and Bias
- Forward Propagation and Derivatives (Forward Pass)
- Backward Propagation and Derivatives (Backward Pass)
- Activation Functions
- Coding vanilla neural network from scratch in python
- Tensorflow deep detailed session
- Making ANN to do housing prices prediction using Tensorflow
- Making ANN to do MNIST digit recognition(60k Images) using Tensorflow
- Introduction to Computer Vision & Image processing
- Introduction to Images in terms of matrices
- Introduction to Filters and Convolution(Non Mathematical manner more into intuition)
- Implementing Convolution on image and applying different filters
- Convolutional Neural Network
- Architecture:
- Padding
- Strided Convolutions
- One Convolutional Layer
- Pooling Layers
- Functions in Tensorflow for CNN
- Coding CNN for Cat & Dog image recognition using core Tensorflow
- Architecture:
- Autoencoders & Variational Autoencoders
- Use cases
- Anomaly Detection
- Image Denoising
- Data Compression
- Architecture
- Introduction to Keras(Derivative of Tensorflow)
- Coding in Tensorflow as well as Keras
- Use cases
- Generative Adversarial Networks: Very fascinating. You can generate your own fake images which have been clicked yet. New Fake images/videos generation
- Deep Learning for Time Series Data, Audio Processing and Natural Language Processing:
- Sequence Models
- Recurrent Neural Network Model
- Different types of Recurrent Neural Networks
- Language model and sequence generation
- Gated Recurrent Unit (GRU)
- Long Short Term Memory (LSTM)
- Deep Recurrent Neural Networks
- Word Representation
- Word embeddings
- Word2Vec
- GloVe word vectors
- Clustering
- K-Means Clustering
- Hierarchical Clustering
- Gaussian Mixture Models
- Data Analysis
- Different types of distribution
- Hypothesis Testing
- Non-Parametric Tests
- ANNOVA
- Learn to choose perfect algorithm for any dataset