Automatic measurement of semantic text similarity is an important task in natural language processing. In this paper, we evaluate the performance of different vector space models to perform this task. We address the real-world problem of modeling patent-to-patent similarity and compare TFIDF (and related extensions), topic models (e.g., latent semantic indexing), and neural models (e.g., paragraph vectors). Contrary to expectations, the added computational cost of text embedding methods is justified only when: 1) the target text is condensed; and 2) the similarity comparison is trivial. Otherwise, TFIDF performs surprisingly well in other cases: in particular for longer and more technical texts or for making finer-grained distinctions between nearest neighbors. Unexpectedly, extensions to the TFIDF method, such as adding noun phrases or calculating term weights incrementally, were not helpful in our context.

Industry is evolving towards Industry 4.0, which holds the promise of increased flexibility in manufacturing, better quality and improved productivity. A core actor of this growth is using sensors, which must capture data that can used in unforeseen ways to achieve a performance not achievable without them. However, the complexity of this improved setting is much greater than what is currently used in practice. Hence, it is imperative that the management cannot only be performed by human labor force, but part of that will be done by automated algorithms instead. A natural way to represent the data generated by this large amount of sensors, which are not acting measuring independent variables, and the interaction of the different devices is by using a graph data model. Then, machine learning could be used to aid the Industry 4.0 system to, for example, perform predictive maintenance. However, machine learning directly on graphs, needs feature engineering and has scalability issues. In this paper we discuss methods to convert (embed) the graph in a vector space, such that it becomes feasible to use traditional machine learning methods for Industry 4.0 settings.

This paper studies a stylized, yet natural, learning-to-rank problem and points out the critical incorrectness of a widely used nearest neighbor algorithm. We consider a model with $n$ agents (users) $\{x_i\}_{i \in [n]}$ and $m$ alternatives (items) $\{y_j\}_{j \in [m]}$, each of which is associated with a latent feature vector. Agents rank items nondeterministically according to the Plackett-Luce model, where the higher the utility of an item to the agent, the more likely this item will be ranked high by the agent. Our goal is to find neighbors of an arbitrary agent or alternative in the latent space. We first show that the Kendall-tau distance based kNN produces incorrect results in our model. Next, we fix the problem by introducing a new algorithm with features constructed from "global information" of the data matrix. Our approach is in sharp contrast to most existing feature engineering methods. Finally, we design another new algorithm identifying similar alternatives. The construction of alternative features can be done using "local information," highlighting the algorithmic difference between finding similar agents and similar alternatives.

Assume we are given a set of items from a general metric space, but we neither have access to the representation of the data nor to the distances between data points. Instead, suppose that we can actively choose a triplet of items (A,B,C) and ask an oracle whether item A is closer to item B or to item C. In this paper, we propose a novel random forest algorithm for regression and classification that relies only on such triplet comparisons. In the theory part of this paper, we establish sufficient conditions for the consistency of such a forest. In a set of comprehensive experiments, we then demonstrate that the proposed random forest is efficient both for classification and regression. In particular, it is even competitive with other methods that have direct access to the metric representation of the data.

A good classification method should yield more accurate results than simple heuristics. But there are classification problems, especially high-dimensional ones like the ones based on image/video data, for which simple heuristics can work quite accurately; the structure of the data in such problems is easy to uncover without any sophisticated or computationally expensive method. On the other hand, some problems have a structure that can only be found with sophisticated pattern recognition methods. We are interested in quantifying the difficulty of a given high-dimensional pattern recognition problem. We consider the case where the patterns come from two pre-determined classes and where the objects are represented by points in a high-dimensional vector space. However, the framework we propose is extendable to an arbitrarily large number of classes. We propose classification benchmarks based on simple random projection heuristics. Our benchmarks are 2D curves parameterized by the classification error and computational cost of these simple heuristics. Each curve divides the plane into a "positive- gain" and a "negative-gain" region. The latter contains methods that are ill-suited for the given classification problem. The former is divided into two by the curve asymptote; methods that lie in the small region under the curve but right of the asymptote merely provide a computational gain but no structural advantage over the random heuristics. We prove that the curve asymptotes are optimal (i.e. at Bayes error) in some cases, and thus no sophisticated method can provide a structural advantage over the random heuristics. Such classification problems, an example of which we present in our numerical experiments, provide poor ground for testing new pattern classification methods.

Time series classification is an increasing research topic due to the vast amount of time series data that are being created over a wide variety of fields. The particularity of the data makes it a challenging task and different approaches have been taken, including the distance based approach. 1-NN has been a widely used method within distance based time series classification due to it simplicity but still good performance. However, its supremacy may be attributed to being able to use specific distances for time series within the classification process and not to the classifier itself. With the aim of exploiting these distances within more complex classifiers, new approaches have arisen in the past few years that are competitive or which outperform the 1-NN based approaches. In some cases, these new methods use the distance measure to transform the series into feature vectors, bridging the gap between time series and traditional classifiers. In other cases, the distances are employed to obtain a time series kernel and enable the use of kernel methods for time series classification. One of the main challenges is that a kernel function must be positive semi-definite, a matter that is also addressed within this review. The presented review includes a taxonomy of all those methods that aim to classify time series using a distance based approach, as well as a discussion of the strengths and weaknesses of each method.

Recently, generative adversarial networks and adversarial autoencoders have gained a lot of attention in machine learning community due to their exceptional performance in tasks such as digit classification and face recognition. They map the autoencoder's bottleneck layer output (termed as code vectors) to different noise Probability Distribution Functions (PDFs), that can be further regularized to cluster based on class information. In addition, they also allow a generation of synthetic samples by sampling the code vectors from the mapped PDFs. Inspired by these properties, we investigate the application of adversarial autoencoders to the domain of emotion recognition. Specifically, we conduct experiments on the following two aspects: (i) their ability to encode high dimensional feature vector representations for emotional utterances into a compressed space (with a minimal loss of emotion class discriminability in the compressed space), and (ii) their ability to regenerate synthetic samples in the original feature space, to be later used for purposes such as training emotion recognition classifiers. We demonstrate the promise of adversarial autoencoders with regards to these aspects on the Interactive Emotional Dyadic Motion Capture (IEMOCAP) corpus and present our analysis.

This paper draws a parallel between similarity-based categorisation models developed in cognitive psychology and the nearest neighbour classifier (1-NN) in machine learning. Conceived as a result of the historical rivalry between prototype theories (abstraction) and exemplar theories (memorisation), recent models of human categorisation seek a compromise in-between. Regarding the stimuli (entities to be categorised) as points in a metric space, machine learning offers a large collection of methods to select a small, representative and discriminative point set. These methods are known under various names: instance selection, data editing, prototype selection, prototype generation or prototype replacement. The nearest neighbour classifier is used with the selected reference set. Such a set can be interpreted as a data-driven categorisation model. We juxtapose the models from the two fields to enable cross-referencing. We believe that both machine learning and cognitive psychology can draw inspiration from the comparison and enrich their repertoire of similarity-based models.

Learning vector embeddings of complex networks is a powerful approach used to predict missing or unobserved edges in network data. However, an open challenge in this area is developing techniques that can reason about $\textit{subgraphs}$ in network data, which can involve the logical conjunction of several edge relationships. Here we introduce a framework to make predictions about conjunctive logical queries---i.e., subgraph relationships---on heterogeneous network data. In our approach, we embed network nodes in a low-dimensional space and represent logical operators as learned geometric operations (e.g., translation, rotation) in this embedding space. We prove that a small set of geometric operations are sufficient to represent conjunctive logical queries on a network, and we introduce a series of increasingly strong implementations of these operators. We demonstrate the utility of this framework in two application studies on networks with millions of edges: predicting unobserved subgraphs in a network of drug-gene-disease interactions and in a network of social interactions derived from a popular web forum. These experiments demonstrate how our framework can efficiently make logical predictions such as "what drugs are likely to target proteins involved with both diseases X and Y?" Together our results highlight how imposing logical structure can make network embeddings more useful for large-scale knowledge discovery.

For statistical learning, categorical variables in a table are usually considered as discrete entities and encoded separately to feature vectors, e.g., with one-hot encoding. "Dirty" non-curated data gives rise to categorical variables with a very high cardinality but redundancy: several categories reflect the same entity. In databases, this issue is typically solved with a deduplication step. We show that a simple approach that exposes the redundancy to the learning algorithm brings significant gains. We study a generalization of one-hot encoding, similarity encoding, that builds feature vectors from similarities across categories. We perform a thorough empirical validation on non-curated tables, a problem seldom studied in machine learning. Results on seven real-world datasets show that similarity encoding brings significant gains in prediction in comparison with known encoding methods for categories or strings, notably one-hot encoding and bag of character n-grams. We draw practical recommendations for encoding dirty categories: 3-gram similarity appears to be a good choice to capture morphological resemblance. For very high-cardinality, dimensionality reduction significantly reduces the computational cost with little loss in performance: random projections or choosing a subset of prototype categories still outperforms classic encoding approaches.