mlpack in Google Summer of Code 2019

In 2019 mlpack applied again for Google Summer of Code. That’s the sixed year in a row and our 5th Summer of Code in total. We received an astounding number of applications and were able to accept 10 talented students. Here is what they will be working on this summer:

Quantum Gaussian Mixture Models

by SangYeon Kim, mentored by Sumedh Ghaisas

SangYeon will implement the Quantum Gaussian Mixture Model (QGMM) for Quantum Clustering, on top of that he will run different experiments to see how fast the method trains, how the method models the data, and search for edge cases and compare it with the classical GMM.

Note

"I like to go short-trip whenever my mind gets stuck with something :)"

Essential Deep Learning Modules

by Toshal Agrawal, mentored by Shikhar Jaiswal

Toshal will provide flexible and extensible implementation of Least Squares Generative Adversarial Networks (LSGANs), Label Smoothing, Bidirectional GAN (BIGAN) and Stacked Generative Adversarial Network (StackGAN). In addition Toshal also aims at providing techniques to measure GAN performance like Inception Score (IS) and Fréchet Inception Distance (FID).

Note

"In my free time I like travelling, reading and playing soccer."

Essential Deep Learning Modules

by Saksham Bansal, mentored by Shikhar Jaiswal

Saksham will implement novel techniques for training GANs efficiently such as mini-batch discrimination, virtual batch normalisation and additionally, adding support for Conditional GAN (CGAN). Future work for Stacked GAN (SGAN) is also proposed.

Note

"In my free time I like watching movies (I recently saw John Wick 3) and reading "books."

Application of ANN Algorithms

by Mehul Kumar Nirala, mentored by Marcus Edel, Sumedh Ghaisas

Mehul will implement various network models using some of the existing building blocks. He will implement VGG 19 for Image Classification, LSTM Networks for Sentiment Analysis, and LSTM for time series prediction.

Note

"I usually watch horror, kind of weird but I like it."

String Processing Utilities

by Jeffin Sam, mentored by Mikhail Lozhnikov

Jeffin will implement String Processing Utilities for boosting mlpack library to manipulate string data types and to convert it into numeric datatype to apply machine learning algorithms.

Note

"I like to watch stand-up comedy in my free time."

Advanced Kernel Density Estimation Improvements

by Roberto Hueso Gomez, mentored by Ryan Curtin

Roberto will work on improvements for the already existing KDE codebase; including improvements in cases where: the data dimensionality is high and the current implementation suffers because of kernel bandwidth scale.

Note

Likes everything related to space.

mlpack-Tensorflow Translator

by Sreenik Seal, mentored by Atharva Khandait

Sreenik will work on a converter that can translate models built in Tensorflow, Keras, Pytorch, Mxnet, ONNX to mlpack's model format and vice-versa. He will also work on displaying the required mlpack code to create a corresponding model.

Note

"I am not a very big fan of movies. I like to play computer games instead."

NeuroEvolution of Augmenting Topologies & Multi-Objective Optimization

by Rahul Ganesh Prabhu, mentored by Marcus Edel

Rahul will implement a NEAT framework as well a framework for multi-objective optimization.

Note

Likes everything around robotics.

"I am not a very big fan of movies. I like to play computer games instead."

Proximal Policy Optimization

by Xiaohong Ji, mentored by Manish Kumar, Marcus Edel

Xiaohong will implement a new family of policy gradient methods for reinforcement learning, which alternate between sampling data through interaction with the environment, and optimizing a "surrogate" objective function using stochastic gradient ascent.

Note

"In my free time, I like to swimming, cooking. Swimming is really a good way to "relax the body."

Particle Swarm Optimization

by Suryoday Basak, mentored by Marcus Edel

Suryoday will implement different variants of particle swarm optimization for ensmallen.

Note

"Since my childhood, I've been a musician - and I'm trained a bit in Western classical as well as Hindustani."

You can find more information on each of the projects on the Summer of Code website.