Need of Systems Approach for Biological Explanation of Anti-Learnable Signatures

Lecture slides:

• Need of System Approach for Biological Explanation of Anti-learnable Signature
• Overview
• Anti-learning in natural data
• CRT response for esophageal cancer
• Prediction of CRT response for Oesophageal Cancer
• Learning and anti-learning mode of supervised classification
• Label permutation test x 1,000; AC data
• Prediction for anti-learnable data
• LTO-cross validation
• KDD’02 task: identification of aryl hydrocarbon receptor genes in yeast
• Classification of KDD’02 data
• Predicting spontaneous termination of paroxysmal atrial fibrillation episodes
• Anti-learning in synthetic data
• Example of perfect anti-learning with linear classifiers
• Geometry of high dimensional samples
• An idea of high-dimensional mimicry

Author: Adam Kowalczyk, National ICT Australia