Learning with structured data
After this course the students will have an in depth understanding of the machine learning techniques used on structured data (input and output). They will be able to successfully apply machine learning algorithms when solving real problems concerned with computational biology, multimedia systems and social networks. They will be able to concept, analyze, realize and evaluate the developed machine learning system performances.
Praktische info:
Leertraject
georganiseerd door:
Course syllabus
Evaluating similarity using kernel, discriminating models, structure modeling using graphical models, connected kernel maps, state vector machines for connected and structured output spaces, efficient max-margin structural classification algorithms, discriminative suffix trees learning for predictions using the perceptron algorithm, search optimization learning, energy based models, generalized limit and structural tagging consistency, Hilbert space, Baes approach to structural models, structural output density measurements. Algorithm application for problems concerned with microorganism characteristics (e.g. temperature resistance, need for oxygen), and phenotype prediction multimedia applications, social media aggregators