Workshop on `Machine Learning in Python’
Mr Dhanesh Solanki, Principal Engineer (Architect), Harman Connected Service Corporation Pvt Ltd (Now Samsung), Bangalore conducted a workshop on `Machine Learning in Python’ on Saturday, 8th February 2020 at Smart Room, Central Block. It was organized by Department of Computer Science, CHRIST (Deemed to be University), Lavasa. Director & Dean, Dr Fr Jossy P George and Academic Coordinator, Dr Fr Arun Antony inaugurated the workshop.
Mr Solanki, holds nearly ten years of experience in Data Science and is currently associated with Harman Connected Service Corporation Pvt. Ltd (Now Samsung) as Principal Engineer (Architect) in Automotive Domain "ADAS". He is also a PhD student at Indian Statistical Institute, Bangalore, and his research topic is Hybrid Deep Neural Network. He holds M Tech from BITS, Pilani and B Tech from University of Rajasthan. He holds six patents which deal with multiple ideas selected from MIT Media Lab Boston Lab Under Harman Research of ADAS Domain. He has submitted many research papers and is privileged to be a reviewer for many international journals.
The workshop began by imparting a simplified understanding about machine leaning, IT, neural network, and deep neural network. Gradually, Mr Dhanesh stepped into the depth of machine learning, when he talked about data mining using machine learning approach, specifically about data preparation constituting data pre-processing which further constitutes data leaning i.e. getting rid of incomplete, noisy and inconsistent data, followed by data integration, data transformation and reduction.
It was enlightening for the students to get hands-on data set and Python code to understand data cleaning and data visualization using python language in addition to some basic mathematical and statistical formulae to perform operations on the dataset provided.
The students were given a brief understanding about decision tree induction inorder to understand the fitness values with the help of real-life examples. The students were also asked to provide a pseudo-code in python as a solution to a problem statement from google maps. This included finding the shortest path from location A to B, from a given number of paths and the distance each covers. The twist was when a landslide occurred in the shortest path. The students enthusiastically participated in the activity.
He further explained feature selection and dimensionality reduction, search strategies and evaluation strategies for building relationships between the data using mutual information, leading to sequential floating backward selection, market-based shopping or weightage probabilities, and dynamic programming.
Last but not the least, Mr Dhanesh made the budding data scientists of CHRIST realise the importance of ‘WHY?’ in Data Science.
Comments
Post a Comment