... is a benchmark in visual object category recognition and detection, providing the vision and machine learning communities with a standard dataset of images and annotation, and standard evaluation procedures. Organised annually from 2005 to present, the challenge and its associated dataset has become accepted as the benchmark for object detection. This paper describes the dataset and evaluation procedure. We review the state-of-the-art in evaluated methods for both classification and detection, analyse whether the methods are statistically different, what they are learning from the images (e.g. the object or its context), and what the methods find easy or confuse. The paper concludes with lessons learnt in the three year history of the challenge, and proposes directions for future improvement and extension.

When humans approach the task of text categorization, they interpret the specific wording of the document in the much larger context of their background knowledge and experience. On the other hand, state-of-the-art information retrieval systems are quite brittle—they traditionally represent documents as bags of words, and are restricted to learning from individual word occurrences in the (necessarily limited) training set. For instance, given the sentence “Wal-Mart supply chain goes real time”, how can a text categorization system know that Wal-Mart manages its stock with RFID technology? And having read that “Ciprofloxacin belongs to the quinolones group”, how on earth can a machine know that the drug mentioned is an antibiotic produced by Bayer? In this paper we present algorithms that can do just that. We propose to enrich document representation through automatic use of a vast compendium of human knowledge—an encyclopedia. We apply machine learning techniques to Wikipedia, the largest encyclopedia to date, which surpasses in scope many conventional encyclopedias and provides a cornucopia of world knowledge. Each Wikipedia article represents a concept, and documents to be categorized are represented in the rich feature space of words and relevant Wikipedia concepts. Empirical results confirm that this knowledge-intensive representation brings text categorization to a qualitatively new level of performance across a diverse collection of datasets.

Abstract—Software defect prediction strives to improve software quality and testing efficiency by constructing predictive classification models from code attributes to enable a timely identification of fault-prone modules. Several classification models have been evaluated for this task. However, due to inconsistent findings regarding the superiority of one classifier over another and the usefulness of metric-based classification in general, more research is needed to improve convergence across studies and further advance confidence in experimental results. We consider three potential sources for bias: comparing classifiers over one or a small number of proprietary data sets, relying on accuracy indicators that are conceptually inappropriate for software defect prediction and cross-study comparisons, and, finally, limited use of statistical testing procedures to secure empirical findings. To remedy these problems, a framework for comparative software defect prediction experiments is proposed and applied in a large-scale empirical comparison of 22 classifiers over 10 public domain data sets from the NASA Metrics Data repository. Overall, an appealing degree of predictive accuracy is observed, which supports the view that metric-based classification is useful. However, our results indicate that the importance of the particular classification algorithm may be less than previously assumed since no significant performance differences could be detected among the top 17 classifiers. Index Terms—Complexity measures, data mining, formal methods, statistical methods, software defect prediction. Ç 1

Label ranking studies the problem of learning a mapping from instances to rankings over a predefined set of labels. Hitherto existing approaches to label ranking implicitly operate on an underlying (utility) scale which is not calibrated in the sense that it lacks a natural zero point. We propose a suitable extension of label ranking that incorporates the calibrated scenario and substantially extends the expressive power of these approaches. In particular, our extension suggests a conceptually novel technique for extending the common learning by pairwise comparison approach to the multilabel scenario, a setting previously not being amenable to the pairwise decomposition technique. The key idea of the approach is to introduce an artificial calibration label that, in each example, separates the relevant from the irrelevant labels. We show that this technique can be viewed as a combination of pairwise preference learning and the conventional relevance classification technique, where a separate classifier is trained to predict whether a label is relevant or not. Empirical results in the area of text categorization, image classification and gene analysis underscore the merits of the calibrated model in comparison to state-of-the-art multilabel learning methods.

Adequate representation of natural language semantics requires access to vast amounts of common sense and domain-specific world knowledge. Prior work in the field was based on purely statistical techniques that did not make use of background knowledge, on limited lexicographic knowledge bases such as WordNet, or on huge manual efforts such as the CYC project. Here we propose a novel method, called Explicit Semantic Analysis (ESA), for fine-grained semantic interpretation of unrestricted natural language texts. Our method represents meaning in a high-dimensional space of concepts derived from Wikipedia, the largest encyclopedia in existence. We explicitly represent the meaning of any text in terms of Wikipedia-based concepts. We evaluate the effectiveness of our method on text categorization and on computing the degree of semantic relatedness between fragments of natural language text. Using ESA results in significant improvements over the previous state of the art in both tasks. Importantly, due to the use of natural concepts, the ESA model is easy to explain to human users. 1.

Abstract. Multilabel classification is an extension of conventional classification in which a single instance can be associated with multiple labels. Recent research has shown that, just like for standard classification, instance-based learning algorithms relying on the nearest neighbor estimation principle can be used quite successfully in this context. However, since hitherto existing algorithms do not take correlations and interdependencies between labels into account, their potential has not yet been fully exploited. In this paper, we propose a new approach to multilabel classification, which is based on a framework that unifies instancebased learning and logistic regression, comprising both methods as special cases. This approach allows one to capture interdependencies between labels and, moreover, to combine model-based and similarity-based inference for multilabel classification. As will be shown by experimental studies, our approach is able to improve predictive accuracy in terms of several evaluation criteria for multilabel prediction. 1

The ability to represent knowledge about space and its position therein is crucial for a mobile robot. To this end, topological and semantic descriptions are gaining popularity for augmenting purely metric space representations. In this paper we present a multi-modal place classification system that allows a mobile robot to identify places and recognize semantic categories in an indoor environment. The system effectively utilizes information from different robotic sensors by fusing multiple visual cues and laser range data. This is achieved using a high-level cue integration scheme based on a Support Vector Machine (SVM) that learns how to optimally combine and weight each cue. Our multi-modal place classification approach can be used to obtain a real-time semantic space labeling system which integrates information over time and space. We perform an extensive experimental evaluation of the method for two different platforms and environments, on a realistic off-line database and in a live experiment on an autonomous robot. The results clearly demonstrate the effec-

Learning from data streams is a research area of increasing importance. Nowadays, several stream learning algorithms have been developed. Most of them learn decision models that continuously evolve over time, run in resource-aware environments, detect and react to changes in the environment generating data. One important issue, not yet conveniently addressed, is the design of experimental work to evaluate and compare decision models that evolve over time. There are no golden standards for assessing performance in non-stationary environments. This paper proposes a general framework for assessing predictive stream learning algorithms. We defend the use of Predictive Sequential methods for error estimate – the prequential error. The prequential error allows us to monitor the evolution of the performance of models that evolve over time. Nevertheless, it is known to be a pessimistic estimator in comparison to holdout estimates. To obtain more reliable estimators we need some forgetting mechanism. Two viable alternatives are: sliding windows and fading factors. We observe that the prequential error converges to an holdout estimator when estimated over a sliding window or using fading factors. We present illustrative examples of the use of prequential error estimators, using fading factors, for the tasks of: i) assessing performance of a learning algorithm; ii) comparing learning algorithms; iii) hypothesis testing using McNemar test; and iv) change detection using Page-Hinkley test. In these tasks, the prequential error estimated using fading factors provide reliable estimators. In comparison to sliding windows, fading factors are faster and memory-less, a requirement for streaming applications. This paper is a contribution to a discussion in the good-practices on performance assessment when learning dynamic models that evolve over time. Categories and Subject Descriptors H.2.8 [Database Management]: Database applications— data mining; I.2.6 [Artificial Intelligence]: Learning—

Quantitative attributes are usually discretized in naive-Bayes learning. We prove a theorem that explains why discretization can be effective for naive-Bayes learning. The use of different discretization techniques can be expected to affect the classification bias and variance of generated naive-Bayes classifiers, effects we name discretization bias and variance. We argue that by properly managing discretization bias and variance, we can effectively reduce naive-Bayes classification error. In particular, we propose proportional k-interval discretization and equal size discretization, two efficient heuristic discretization methods that are able to effectively manage discretization bias and variance by tuning discretized interval size and interval number. We empirically evaluate our new techniques against five key discretization methods for naive-Bayes classifiers. The experimental results support our theoretical arguments by showing that naive-Bayes classifiers trained on data discretized by our new methods are able to achieve lower classification error than those trained on data discretized by alternative discretization methods.

Reliably predicting software defects is one of the holy grails of software engineering. Researchers have devised and implemented a plethora of defect/bug prediction approaches varying in terms of accuracy, complexity and the input data they require. However, the absence of an established benchmark makes it hard, if not impossible, to compare approaches. We present a benchmark for defect prediction, in the form of a publicly avail-able dataset consisting of several software systems, and provide an extensive com-parison of well-known bug prediction approaches, together with novel approaches we devised. We evaluate the performance of the approaches using different perfor-mance indicators: classification of entities as defect-prone or not, ranking of the entities, with and without taking into account the effort to review an entity. We performed three sets of experiments aimed at (1) comparing the approaches across different systems, (2) testing whether the differences in performance are statistically significant, and (3) investigating the stability of approaches across different learners. Our results indicate that, while some approaches perform better than others in a statistically significant manner, external validity in defect prediction is still an open problem, as generalizing results to different contexts/learners proved to be a partially unsuccessful endeavor.