- 15 students
- 82 lessons
- 0 quizzes
- 10 week duration
Machine Learning is rapidly developing field. Over 50,000 jobs are vacant in India and the count is only going to increase with time. These jobs are not getting filled because of lack of talented engineers who can apply machine learning to various data sets. It must be appreciated that Data Science is a demanding subject; it is in fact culmination of Statistics, Linear Algebra, Calculus and Computer Programming. This doesn’t mean that cracking data science is impossible! It only means that we need a quick introduction to data science where Statistics, Linear Algebra, Calculus and Computer programming is taught simultaneously as a hands-on session and a follow up with regular classes with focus on real case studies. All these concepts are subjects which require the student to develop an intuition towards observing real world data and correlate it with models and algorithms of machine learning. Since it is an intense and demanding task to master this subject, the initial foundation must be strong. Our course is created to create continuous discussion and meditation on the concepts to help students to find a foot-hold on this rapidly advancing field.
1.1 Python is the language we use to tell our computers to compute a set of algorithms: why Python?
Imagine a real python (snake); it’s big, wavy and long. We want something similar to handle large number of related/unrelated data sets. A python (snake) is continuous, flexible and similarly Python (language) provides us with that flexibility to handle tabulated/listed arrays of data. Since machine learning at advanced level is all about dealing with software implementation of linear algebra and multinomial expression(not complicated!) , it is important to choose a language that suits easy handling of grouped and structured data.
1.2 Machine Learning:
Mathematics related to computational science provides us with many techniques to discover hidden patterns in data. And because of advanced programming languages (advanced in sense of handling abstract concepts and data structures) like python , Scala etc we have the resource to rapidly code algorithms or to implement it using readily available libraries . Every problem is different because every set of data have different quality. This field and has evolved over 50 years and derives its concreteness from Statistics.
This is all about numbers. It is a field of math which centers on techniques to discover how much truth is hidden in given data set. Data may show something but the reality is usually hidden deep in the distributions. Once this data is plotted on to a graph we begin to see a pattern. This pattern gives us the insight to go ahead and take the next necessary steps. A statistician is the one who is trained to see Data-Set as if it was a physical entity. That’s why he/she uses terms like Gradient, Descent, skew, etc. Statistics is all about making error free interpretations about the phenomenon represented by the data. Many tools have been developed over years. We will learn many (important ones) in the workshop and code it in python simultaneously.
But one cannot assume that mere understanding of algorithms can help us solve any prediction problem!
If that was true, then we could have used “Machine learning algorithms” to teach itself “Machine Learning” and then we would have Evolving Algorithms which are found in science fiction movies. However, the reality is different. Imagine a carpenter and his toolbox. He has a hammer, a set of pliers, nails, files, cutters, hacksaw and many others handy tools. We all know how each one of them functions. Even then, it still does not make us Carpenters! It is only a carpenter who knows which tool to use in a certain situation.
“Data Engineering” is the ability to implement machine learning algorithms, and
“Data Science” is knowing which algorithm to use, when and where.
Data science is an Art. This art can be developed when one develops an intuition for data. Diploma in Machine Learning and AI is tailored such that the student develops this intuition.
Introduction To Machine Learning
- What is Machine learning and real world examples
- Machine, its language and evolution into AI
- A short trip on evolution of mathematics and science.
- Supervised learning
- Unsupervised learning
- Model representation
- Cost function & intuition
- Gradient decent & Intuition
- Gradient decent for linear regression intuition
- Classification and Representation
- Revisiting cost function
- Regularizing linear regression
- Evaluation a learning algorithm
LIBRARIES AND API'S IN PYTHON
STATISTICS & PROBABILITY USING PYTHON
- Types of data
- Mean, Median , Mode using python
- Standard deviation using python
- Probability Density function and Probability mass function
- Common data distribution
- Hands On Activity Covariance and Correlation
- Probability refresher
- Problem on Conditional Probability
- Solution Conditional Probability
- Baye’s Theorem
MACHINE LEARNING AND PYTHON
- Supervised vs. Unsupervised Learning, and Train/Test
- Linear Regression
- Polynomial Regression
- Logistic Regression
- Multivariate Analysis: Predict housing Prices
- Preventing Overfitting
- Bayesian Methods
- Implementing a Spam Classifier with Naive Bayes
- K-Means Clustering
- Hands On Activity Clustering people based on income and Education
- Measuring the chaos in data: Entropy
- Decision Trees
- Project: Apply Decision Trees concepts
- Ensemble Learning
- Meditations on reference systems (Cartesian, spherical etc.) and linear algebra
- Support Vector Machines (SVM)
- Project: Apply SVM concepts using Scikit
- Principal Component Analysis (PCA) and Dimensionality Reduction
- PCA example with the iris data set
- Percentiles and Moments.
- Data Warehousing: ETL and ELT concepts
- Project using matplotlib
RECOMMENDER SYSTEM: HOW NETFLIX WORKS
DEALING WITH REAL WORLD DATA
APACHE SPARK, ML AND BIG DATA
DESIGNING AN DATA SCIENCE EXPERIMENT
ADVANCED TECHNIQUES: NEURAL NETWORKS AND DEEP LEARNING
- Neural Network Game
- Deep Learning concepts
- Biological Motivation of Neural Networks and Historical
- Deep Learning using Tensoflow on google tensor flow playground
- Tensor Flow
- Convolution Neural Networks(CNN’s)
- CNN for pattern recognition in hand writing recognition
- Recurrent Neural Networks(RNN)
- Sentiment Analysis.
4.56 average based on 16 ratings
This course was really awesome with latest course materials and syllabus . I would recommend it for anyone interested in Machine Learning and AI
Really enjoyed the course .The teaching was made very interesting by the tutor and very clearly articulated
Lots of useful exercises in the course to give practical knowledge . Must for anyone interested in Machine Learning and AI
Highly recommend to people who like programming. All lectures are explained beutifully
Thank you for the course. Learned many new things . The exercises and challenges after each lecture made my basics strong.