Multi-label classification is the task of predicting potentially multiple labels for a given instance. This is common in several applications such as image annotation, document classification and gene function prediction. In this paper we present a formulation for this problem based on reverse prediction: we predict sets of instances given the labels. By viewing the problem from this perspective, the most popular quality measures for assessing the performance of multi-label classification admit relaxations that can be efficiently optimised. We optimise these relaxations with standard algorithms and compare our results with several stateof-the-art methods, showing excellent performance. 1

We pose transductive classification as a matrix completion problem. By assuming the underlying matrix has a low rank, our formulation is able to handle three problems simultaneously: i) multi-label learning, where each item has more than one label, ii) transduction, where most of these labels are unspecified, and iii) missing data, where a large number of features are missing. We obtained satisfactory results on several real-world tasks, suggesting that the low rank assumption may not be as restrictive as it seems. Our method allows for different loss functions to apply on the feature and label entries of the matrix. The resulting nuclear norm minimization problem is solved with a modified fixed-point continuation method that is guaranteed to find the global optimum. 1

This volume contains research papers accepted for presentation at the 1st International Workshop on Learning from Multi-Label Data (MLD’09), which will be held in Bled, Slovenia, at September 7, 2009 in conjunction with ECML/PKDD 2009. MLD’09 is devoted to multi-label learning, which is an emerging and promising research topic of machine learning. In multi-label learning, each example is associated with multiple labels simultaneously, which therefore encompasses traditional supervised learning (single-label) as its special case. Multi-label learning is related to various machine learning paradigms, such as classification, ranking, semi-supervised learning, active learning, multi-instance learning, dimensionality reduction, etc. Initial attempts on multi-label learning date back to 1999 with works on multi-label text categorization. In recent years, the task of learning from multi-label data has been addressed by a number of methods adapted from various popular learning techniques, such as neural networks, decision trees, k-nearest neighbors, kernel methods, ensemble methods, etc. More impressively, multi-label learning has manifested its effectiveness in a diversity of real-world applications, such as image/video annotation, bioinformatics,

In this paper we present an algorithm to learn a multi-label classifier which attempts at directly optimising the F-score. The key novelty of our formulation is that we explicitly allow for assortative (submodular) pairwise label interactions, i.e., we can leverage the co-ocurrence of pairs of labels in order to improve the quality of prediction. Prediction in this model consists of minimising a particular submodular set function, what can be accomplished exactly and efficiently via graph-cuts. Learning however is substantially more involved and requires the solution of an intractable combinatorial optimisation problem. We present an approximate algorithm for this problem and prove that it is sound in the sense that it never predicts incorrect labels. We also present a nontrivial test of a sufficient condition for our algorithm to have found an optimal solution. We present experiments on benchmark multi-label datasets, which attest the value of the proposed technique. We also make available source code that enables the reproduction of our experiments. 1

This volume contains the research papers that form the programme of the 2nd International Workshop on Learning from Multi-Label Data (MLD’10), which was held in

In real world multilabel problems, it is often the case that e.g. documents are simultaneously classified with labels from multiple domains, such as genres in addition to topics. In practice, each of these problems is solved independently without taking advantage of possible label correlations between domains. Following the multi-task learning setting, in which multiple similar tasks are learned in parallel, we propose a global learning approach that jointly considers all domains. It is empirically demonstrated in this work that this approach is effective despite its simplicity when using a multilabel learner that takes label correlations into account. 1.

Researchers have used topic modeling and concept location to understand the latent topics of software development artifacts. These techniques use unsupervised machine-learning algorithms to recover topics. These topics are word-lists and are difficult to distinguish and interpret. Topics are not meaningful until they have been named or interpreted. Current topic labelling approaches are manual, and do not use domain-specific knowledge to improve, contextualize, or describe results for the developers. We propose a solution: labelled topic extraction. Topics are extracted using Latent Dirichlet Allocation (LDA) from commit-log comments recovered from source control systems such as CVS and Bit-Keeper. These topics are given labels relating to a generaliz-CategoriesandSubjectDescriptors

This volume contains research papers accepted for presentation at the 1st International Workshop on Learning from Multi-Label Data (MLD’09), which will be held in Bled, Slovenia, at September 7, 2009 in conjunction with ECML/PKDD 2009. MLD’09 is devoted to multi-label learning, which is an emerging and promising research topic of machine learning. In multi-label learning, each example is associated with multiple labels simultaneously, which therefore encompasses traditional supervised learning (single-label) as its special case. Multi-label learning is related to various machine learning paradigms, such as classification, ranking, semi-supervised learning, active learning, multi-instance learning, dimensionality reduction, etc. Initial attempts on multi-label learning date back to 1999 with works on multi-label text categorization. In recent years, the task of learning from multi-label data has been addressed by a number of methods adapted from various popular learning techniques, such as neural networks, decision trees, k-nearest neighbors, kernel methods, ensemble methods, etc. More impressively, multi-label learning has manifested its effectiveness in a diversity of real-world applications, such as image/video annotation, bioinformatics,

topic naming to support cross-project analysis

Traditional error-correctingoutput codes (E-COCs) decompose a multi-class classification problem into many binary problems. Although it seems natural to use ECOCs for multi-label problems as well, doing so naively createsissues related to: the validity of the encoding, the efficiency of the decoding, the predictabilityofthegeneratedcodeword,and the exploitation of the label dependency. Using canonical correlation analysis, we propose an error-correcting code for multi-label classification. Labeldependencyischaracterized as the most predictable directions in the label space, which are extracted as canonical output variates and encoded into the codeword. Predictions for the codeword define a graphical model of labels with both Bernoulli potentials (from classifiers on the labels) and Gaussian potentials (from regression on the canonical output variates). Decoding is performed by mean-field approximation. We establish connections between the proposed code and research areas such as compressed sensing and ensemble learning. Some of these connections contribute to better understanding of the new code, and others lead to practical improvements in code design. In our empirical study, the proposed code leads to substantial improvements compared to various competitors in music emotion classification and outdoor scene recognition. 1