INIT/AERFAI Summer School on MACHINE LEARNING Beniccasim, Spain June 2013 Valencia Lectures ● Session 1 - Learning Theory Dr. Gábor Lugosi – University Pompeu Fabra (Spain) ● Session 2 - Learning with Multiple Classifier Systems Dr. Gavin Brown – University of Manchester (UK) ● Session 3 - Kernel Methods for Learning Prof. John Shawe-Taylor – University College London (UK) ● Session 4 - Dissimilarity Representation for Classification Prof. Robert P.W. Duin – Delft University of Technology (The Netherlands) Lectures (2) ● Session 5 - Transfer Learning and Adversarial Learning Prof. Dr. Tobias Scheffer – University of Potsdam (Germany) ● Session 6 - Learning from Streaming Data Dr. João Gama – University of Porto (Portugal) ● Session 7 - ROC Analysis and Performance Evaluation Metrics Prof. Peter A. Flach – University of Bristol (UK) ● Session 8 - Statistical Analysis of Experiments Prof. Francisco Herrera – University of Granada (Spain) Session 1 - Learning Theory ● Quick introduction of ML ● Empirical risk minimization ● Inequalities ● Random projections ● ...and much more math... Session 2 - Learning with Multiple Classifier Systems ● Combining voters – estimating possible error ● Sequential (voting) – Bagging : Bootstrap AGGregatING – Random/Rotation forests ● Parallel combination (fix error of predictor) – Boosting Session 3 - Kernel Methods for Learning ● ...so much non trivial math :-/ Session 4 - Dissimilarity Representation for Classification ● Not what is common in data but what is different ● Distances Session 5 - Transfer Learning and Adversarial Learning ● ...looking for smart students :) ● Taking drugs – overlap ● Bayesian methods ● Minimizing risk ● SVN – finding minimum Session 6 - Learning from Streaming Data ● ...finally presentation for humans :) ● Huge amount of data in non-static world ● Incorporating, detecting changes, forgetting ● Landmark / Sliding Window ● Top-K elements problem ● Air quality monitoring – rising an alarm when condition exceeds limits Session 7, Session 8 - Statistical Analysis of Experiments ● Confidence bands ● Testing hypotheses ● Correlation ● Enroll in subjects Statistics I, II ● ...I was lying on the beach that time ;-) Thank You for your attention! Peñíscola