Abstract

We hope to introduce the wonderful world of random forests. Random forests are one of the most successful ensemble methods in machine learning with state-of-the-art performance in many application domains. It works by averaging several predictions of de-correlated trees. In this talk, we will present a unifying view of random forest that is capable of solving almost all your learning problems: classification, regression, density estimation, manifold learning, and semi-supervised learning. We will also discuss recent work on the mathematical forces behind the success of random forest. Several future directions will be mentioned.

References: 1) A. Criminisi, J. Shotton, E. Konukoglu, Decision forests: A unified framework for classification, regression, density estimation, manifold learning and semi-supervised learning, Foundations and Trends in Computer Graphics and Vision, 2012 (available online as an MSR techinal report) 2) G. Biau, L. Devroye, and G. Lugosi, Consistency of Random Forests and Other Averaging Classifiers, JMLR , 2008 3) G. Biau, Analysis of a random forests model, JMLR , 2012

The RCC will be tutorial in nature, and will assume no prior knowledge, however, familiarisation with the notation in Section 2 of [1] would be useful, and for those more theoretically inclined we recommend reading [3] as we shall skip most of the mathematical details.