Google Summer of Code 2016 Projects

Google Summer of Code 2016

Welcome Students (and mentors)! See how to get involved and our GSoC blog posts.

Potential projects

This is a growing list. To add, please create a new wiki page for each project that you describe. Name them as "GSoC_2016_project_XXX" etc. Here is a template.

Main focus

This year's GSoC is about improving Shogun, rather than extending it (exceptions allowed). We also want to recruit new long-term developers.

• Fewer new algorithms. Rather improve existing ones: easier use, efficiency, better documentation, more applications
• Fewer students. A chance for more intense mentoring, interaction between students, blogging, documenting

2016 Mentors

If you have questions about GSoC 2016, ask Lea

Projects

Bold projects are more likely to happen. ALL students will be required to:

• peer-review a fellow student's work in the final phase of GSoC
• jointly helpo with the 5.0 release in the final phase of GSoC

Furthermore, we plan an experiment this year: an additional side-project that all students work on jointly and in addition to their main project -- building a Shogun cookbook

Improving Shogun

These are roughly ordered in our priority in them. Most of them do not focus on Machine Learning but rather on software engineering.

• Easy installation on major platforms
• Unified ML interface, plugin-based architecture
• Fundamental ML: The usual suspects
• A Shogun Detox
• SWIG, Matlab & modular interfaces
• HMM cleanup and application
• Native MS Windows port
• Unifying Shogun's linear algebra
• Flexible modelselection 2

Extending Shogun:

The projects we would like to limit in numbers.

Algorithms

• Gaussian Processes & tensorflow autodiff
• Hip Deep learning
• Fundamental ML: LGSSMs
• Density Estimation in Infinite Dimensional Exponential Families
• Large scale statistical testing
• Solver for the KKT System
• Dual coordinate ascent solver for SO-SVM
• LP/QP Framework
• Debiasing & Cluster computing

Framework

• MCMC & Stan
• Independent jobs Framework
• Shogun cloud extensions

Other ideas:

• Cool pipelines
  • A kaggle pipeline for supervised prediction.
  • Spectrometer (there is an open-source hardward project on this)
  • Music brainz predictions (The cool hair guy at GSoC is the one we should talk to here)
  • Some bio thing?
  • Collaboration with MLPack for toolkit wide performance/accuracy testing

Infrastructure:

• REST interface