Resource Person on Machine Learning
This is an abstract of an event where I was invited as Resource Person on workshop on Machine Learning held at Maharaja Institute of Technology, Thandavapura (Mysore).
What is Artificial Intelligence?
Artificial Intelligence is the science of intelligent programs based on data.
As they say, Artificial Intelligence (AI) the Last Invention We Will Ever Make. But to have a working artificially intelligent machines, the machine needs to learn as we do.
What is Machine Learning?
To make machines learn, the traditional programming paradigm won’t work efficiently. Hence, we need a new way of making machines learn. This new approach to programming is called “Machine Learning”.
Machine Learning is a scientific study of algorithms and statistical models that make systems to automatically learn and improve from experience without being explicitly programmed.
To make ‘Machines’ learn without being programmed explicitly, we need algorithms, mathematical and statistical concepts. Some of them are as mentioned below:
- Linear algebra
- Probability theory-Probability and Statistics
- Graph theory
- Optimization methods
In my sessions at the venue, the talk was focused on hands-on sessions for final year students. First students were taught Machine Learning basics using TensorFlow and other tools.
Hello World in Machine Learning
The first session was involved in training a model to perform a prediction(Hello World) in Google Colabs using python programming language. In this session, students learnt the importance of the following concepts:
- Why Keras is used?
- How to define layers?
- What is optimizer?
- How errors are calculated?
- Why define Epochs?
- Why use numpy?
The next session involved discussions on various supervised and unsupervised learning models.
- Bayesian network considering medical data.
- Expectation–maximization(EM) algorithm vs K means using the iris data set
- naïve Bayesian classifier
- ANN Backpropagation
On a final note, the importance of training models using GPU over CPU was discussed. Added to previous discussions various public and open datasets were explored. Thanks to N N Naveen for his company throughout the day.