mlpack in Google Summer of Code 2018

This year is shaping up to be another productive and exciting year for mlpack; after being accepted into Google Summer of Code 2018, we received an astounding number of applications and were able to accept 6 talented students. Here is what they will be working on this summer:

Implementing Essential Deep Learning Modules

by Shikhar Jaiswal, mentored by Marcus Edel

Shikhar will pick up and extend the pieces from last year's GSoC and will work on implementing Generative adversarial networks (GANs); enriched with a bunch of neat features like Transposed Convolution and validate everything with some really cool experiments.

Shikhar is a second-year undergraduate student pursuing a Bachelors of Technology in Computer Science and Engineering at Indian Institute of Technology, Patna, India.

Note

"If you fancy foreign made films, I'd suggest some k-movies, Oldboy, Train to Busan and I Saw The Devil (watch the original with English subtitles). Though I must warn, they are not for the squeamish."

Neural Collaborative Filtering

by Haritha S. Nair, mentored by Marcus Edel

Haritha will extend the existing collaborative filtering methods to support neural collaborative filtering, which uses neural networks to represent complex user-item interactions and can handle implicit feedback.

Haritha is a 3rd-year Information Technology student, currently undergoing the final year of B.Tech, Atal Bihari Vajpayee - Indian Institute of Information Technology and Management, Gwalior.

Note

"I love reading, mainly thrillers and mystery, currently having my annual harry potter marathon."

Automated Java Binding Generator

by Yasmine Dumouchel, mentored by Kevin Avignon and Ryan Curtin

Yasmine will significantly improve the automatic bindings system we have in place by writing automatically-generated Java bindings. This would allow Java users to have access to the library while keeping the system maintenance burden low and making addition of new bindings even more efficient.

Yasmine went to McGill University in Canada for her undergrad and will pursue his Masters next year at the Université de Montréal.

Note

"When its grey and gloomy I watch/rewatch Silicon Valley or Rick and Morty (which I highly recommend if you have not watched it yet!!)"

Variational Autoencoders

by Atharva Khandait, mentored by Sumedh Ghaisas

Atharva will design and develop a generic class to create variational autoencoders; specifically, this class will integrate into the existing codebase and will work both with feedforward and recurrent neural networks. In addition, he will work on Beta VAE, Conditional Modelling and Regularization.

Atharva is an undergraduate student pursuing a Bachelors degree in Aerospace Engineering at Indian Institute of Technology, Bombay.

Note

"Also like to watch standup comedy, it's a relatively new interest and I have never been to an actual live show, just videos [...] my favourite is Louis CK"

LMNN (via LRSDP) & BoostMetric Implementation

by Manish Kumar, mentored by Ryan Curtin

Manish will first of all work on implementing LMNN on the grounds of LRSDP solver & primal-dual interior point solver and afterwards, he will take a look at BoostMetric which combines boosting and metric learning approaches into one package. In addition, benchmarking will be done to verify that mlpack's implementations are competitive with other implementations of these algorithms.

Manish is a 3rd-year undergrad at IIT (ISM) Dhanbad, India, Applied Mathematics Dept.

Note

"Nothing beats Rick and Morty=D"

Alternatives to Neighborhood-Based Collaborative Filtering

by Wenhao Huang, mentored by Mikhail Lozhnikov

Wenhao will work on improving the existing CF framework, provide better rating predictions, fast executions; packaged in a flexible API. Specifically, Wenhao will work on data normalization, aggregation of neighbor ratings, BiasSVD, SVD++. He will also benchmark the additions against public datasets like MovieLens and Netflix.

Wenhao is currently a senior student at Peking University, Beijing; next semester he will pursue a Master Degree in Computer Science at Carnegie Mellon University.

Note

"I am a food lover"

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

Anyway, congratulations to Shikhar, Haritha, Yasmine, Atharva, Manish and Wenhao! The coding period starts May 14.