Therefore, diagnostic algorithms must reason backwards from symptoms to causes. For example, determining that a dead battery is the cause of your car not starting in the morning (and not the wiring or the ignition switch). The domains of diagnostic algorithms includes analog and digital circuits, software systems, thermal systems, biological systems, and physical mechanisms. The same classes of diagnostic algorithms can apply in all domains. Diagnostic algorithms make observations, often in real time, of a system being diagnosed.

Ecological systems are inherently complex. The processes that affect the distributions of animals and plants operate at multiple spatial and temporal scales, presenting a unique challenge for the development and coordination of effective conservation strategies, particularly for wide-ranging species. In order to study ecological systems across scales, data must be collected at fine resolutions across broad spatial and temporal extents. Crowdsourcing has emerged as an efficient way to gather these data by engaging large numbers of people to record observations. However, data gathered by crowdsourced projects are often biased due to the opportunistic approach of data collection. In this article, we propose a general class of models called AdaSTEM, (for adaptive spatio-temporal exploratory models), that are designed to meet these challenges by adapting to multiple scales while exploiting variation in data density common with crowdsourced data. To illustrate the use of AdaSTEM, we produce intra-seasonal distribution estimates of long-distance migrations across the Western Hemisphere using data from eBird, a citizen science project that utilizes volunteers to collect observations of birds. Subsequently, model diagnostics are used to quantify and visualize the scale and quality of distribution estimates. This analysis shows how AdaSTEM can automatically adapt to complex spatiotemporal processes across a range of scales, thus providing essential information for full-life cycle conservation planning of broadly distributed species, communities, and ecosystems.

In order to emphasize the importance of implemented systems for the field, we also called for system demonstrations; 7 were accepted for the conference, 1 of them associated with a research abstract and 6 of them described in a demonstration extended abstract. At this edition of ICAIL, the Donald H. Berman best student paper award was won by Tran Thi Oanh (Japan Advanced Institute of Science and Technology; JAIST) for the paper entitled "Reference Resolution in Legal Texts" that she wrote with Minh Le Nguyen and Akira Shimazu. Traditionally, ICAIL hosts a lively and varied program of tutorials and workshops. At this conference, there were tutorials covering an introduction to artificial intelligence and law, web ontology and data design, LegalRuleML, and textual information extraction. There were workshops on argumentation, coherence, open and smart data, evidence, e-discovery, e-justice, and network analysis. Also, the international workshop series, Computational Models of Natural Argument, joined ICAIL for its 13th edition (CMNA XIII). The conference was held under the auspices of the Senate of the Italian Republic with as hosting institution the Consiglio Nazionale delle Ricerche (National Research Council of Italy), central unit in Rome. Both AAAI and ACM SIGART were in cooperation. Conference officials were Bart Verheij (program chair), Enrico Francesconi (conference chair), and Anne Gardner (secretary/treasurer).

Recommendation systems have been integrated into the majority of large online systems. They tailor those systems to individual users by filtering and ranking information according to user profiles. This adaptation process influences the way users interact with the system and, as a consequence, increases the difficulty of evaluating a recommendation algorithm with historical data (via offline evaluation). This paper analyses this evaluation bias and proposes a simple item weighting solution that reduces its impact. The efficiency of the proposed solution is evaluated on real world data extracted from Viadeo professional social network.

AUC (area under ROC curve) is an important evaluation criterion, which has been popularly used in many learning tasks such as class-imbalance learning, cost-sensitive learning, learning to rank, etc. Many learning approaches try to optimize AUC, while owing to the non-convexity and discontinuousness of AUC, almost all approaches work with surrogate loss functions. Thus, the consistency of AUC is crucial; however, it has been almost untouched before. In this paper, we provide a sufficient condition for the asymptotic consistency of learning approaches based on surrogate loss functions. Based on this result, we prove that exponential loss and logistic loss are consistent with AUC, but hinge loss is inconsistent. Then, we derive the $q$-norm hinge loss and general hinge loss that are consistent with AUC. We also derive the consistent bounds for exponential loss and logistic loss, and obtain the consistent bounds for many surrogate loss functions under the non-noise setting. Further, we disclose an equivalence between the exponential surrogate loss of AUC and exponential surrogate loss of accuracy, and one straightforward consequence of such finding is that AdaBoost and RankBoost are equivalent.

In numerous applicative contexts, data are too rich and too complex to be represented by numerical vectors. A general approach to extend machine learning and data mining techniques to such data is to really on a dissimilarity or on a kernel that measures how different or similar two objects are. This approach has been used to define several variants of the Self Organizing Map (SOM). This paper reviews those variants in using a common set of notations in order to outline differences and similarities between them. It discusses the advantages and drawbacks of the variants, as well as the actual relevance of the dissimilarity/kernel SOM for practical applications.

In this paper we provide a new analysis of the SEM algorithm. Unlike previous work, we focus on the analysis of a single run of the algorithm. First, we discuss the algorithm for general mixture distributions. Second, we consider Gaussian mixture models and show that with high probability the update equations of the EM algorithm and its stochastic variant are almost the same, given that the input set is sufficiently large. Our experiments confirm that this still holds for a large number of successive update steps. In particular, for Gaussian mixture models, we show that the stochastic variant runs nearly twice as fast.

Despite the major advances taken in causal modeling, causality is still an unfamiliar topic for many statisticians. In this paper, it is demonstrated from the beginning to the end how causal effects can be estimated from observational data assuming that the causal structure is known. To make the problem more challenging, the causal effects are highly nonlinear and the data are missing at random. The tools used in the estimation include causal models with design, causal calculus, multiple imputation and generalized additive models. The main message is that a trained statistician can estimate causal effects by judiciously combining existing tools.

The learning with privileged information setting has recently attracted a lot of attention within the machine learning community, as it allows the integration of additional knowledge into the training process of a classifier, even when this comes in the form of a data modality that is not available at test time. Here, we show that privileged information can naturally be treated as noise in the latent function of a Gaussian Process classifier (GPC). That is, in contrast to the standard GPC setting, the latent function is not just a nuisance but a feature: it becomes a natural measure of confidence about the training data by modulating the slope of the GPC sigmoid likelihood function. Extensive experiments on public datasets show that the proposed GPC method using privileged noise, called GPC+, improves over a standard GPC without privileged knowledge, and also over the current state-of-the-art SVM-based method, SVM+. Moreover, we show that advanced neural networks and deep learning methods can be compressed as privileged information.

Heuristic functions based on the delete relaxation compute upper and lower bounds on the optimal delete-relaxation heuristic h+, and are of paramount importance in both optimal and satisficing planning. Here we introduce a principled and flexible technique for improving h+, by augmenting delete-relaxed planning tasks with a limited amount of delete information. This is done by introducing special fluents that explicitly represent conjunctions of fluents in the original planning task, rendering h+ the perfect heuristic h* in the limit. Previous work has introduced a method in which the growth of the task is potentially exponential in the number of conjunctions introduced. We formulate an alternative technique relying on conditional effects, limiting the growth of the task to be linear in this number. We show that this method still renders h+ the perfect heuristic h* in the limit. We propose techniques to find an informative set of conjunctions to be introduced in different settings, and analyze and extend existing methods for lower-bounding and upper-bounding h+ in the presence of conditional effects. We evaluate the resulting heuristic functions empirically on a set of IPC benchmarks, and show that they are sometimes much more informative than standard delete-relaxation heuristics.