Fernando J. Iglesias García

Contact

Email: [email protected] irc: iglesiasg

Personal

Occupation: Marie Curie research fellow at Thales Nederland and PhD student at the University of Twente, associated with the faculty of EEMCS. Research interests: statistical signal processing and information fusion. Mainly, the application of Bayesian statistics and Monte Carlo methods to target tracking. Other interests: machine learning in education and open source, the amazing evolution of robotics, competitive programming as a tool to improve problem solving skills, fitness, travel, and video games.

Shogun

Fernando got involved in Shogun back at the beginning of 2012 with the goal of becoming a GSoC student, which he did, starting Shogun's framework for structured output learning under the supervision of Nico Görnitz and Georg Zeller.

At the end of 2012 and during the first half of 2013, Fernando used Shogun for his undergraduate thesis work on structured output learning applied to label sequence learning and learning of general structured output models (aka graphs). The latter part was the precursor of GSoC 2013 project developed by @hushell and mentored by Patrick Pletscher.

During the summer of 2013, Fernando implemented in Shogun the metric learning algorithm known as Large Margin Nearest Neighbours (LMNN) and applied it to a metagenomics data set. This work was performed under the supervision of Georg Zeller.

Ask him about

• The primal SO-SVM and its implementation in Shogun.
• k-NN and LMNN.
• Using valgrind to catch memory leaks and profiling Shogun code.
• Anything else :-)