Google Summer of Code
30 May 2017
Hey guys!
The community bonding period is over, and I am fully charged and psyched to start coding! :D
Before actually starting/knwoing about the Community bonding, I thought why are these guys setting 1 month just for interactions and setup.
Guess its a really important part of the entire GSoC timelines, as you get to interact with developers from your mentoring organisation and get up to speed with what has already been developed related to your project.
Machine learning is really being picked up by every computational-mathematical-RnD-financial domain based teams around the world. SCILAB is infact taking the open source platform for numerical computation higher up the tech-hill, and from here on I think its going to be only Gradient Ascent for the Scilab team. :D
Though machine learning has picked up pace in the last 4-5 years, the work around it has been going since a long duration. Scilab being no different has its own set of toolboxes’ for machine learning. The issue is that these all are scattered instances of a particular machine learning algorithm - regression, clustering, neural nets. Eventually it would become essential for Scilab to have an unified machine learning toolbox which encompasses all vital machine learning algorithms.
So during the community bonding period I reinforced my knowledge and experience working with famous Ml libraries based in python.
| Tasks | Description | Status |
|---|---|---|
| Getting to know my mentors and finalizing a rough starting point for the project | I was asked to revise all major machine learning models/ algorithms and get clarity about their mathematical modeling | Done |
| Getting in-depth knowledge of scikit-learn library | I went through the documentation and tutorials on the scikit-learn portal. Got hands-on experience working with the various modules in sklearn and trying them on Kaggle datasets | Done |
| Get acquainted with deep learning libraries like tensorflow and keras | Since neural network module already exists in scilab, it is current need to have a deep learning module/library implementation through Scilab. I understood the working and modeling of neural networks using tensorflow and keras library for python | Done |
| Scilab syntaxes and toolboxes | Here I was required to make myself comfortable in using Scilab for the entire development period. Getting to know how and where Scilab is different from MATLAB(which I am used to) | Done |
Let the work begin :D