Search

Reservoir Computing (RC) models, a subclass of recurrent neural networks, are distinguished by their fixed, non-trainable input layer and dynamically coupled reservoir, with only the static readout layer being trained. This design circumvents the issues associated with backpropagating error signals through time, thereby enhancing both stability and training efficiency. RC models have been successfully applied across a broad range of application domains. Crucially, they have been demonstrated to be universal approximators of time-invariant dynamic filters with fading memory, under various settings of approximation norms and input driving sources. Simple Cycle Reservoirs (SCR) represent a specialized class of RC models with a highly constrained reservoir architecture, characterized by uniform ring connectivity and binary input-to-reservoir weights with an aperiodic sign pattern. For linear reservoirs, given the reservoir size, the reservoir construction has only one degree of freedom -- the reservoir cycle weight. Such architectures are particularly amenable to hardware implementations without significant performance degradation in many practical tasks. In this study we endow these observations with solid theoretical foundations by proving that SCRs operating in real domain are universal approximators of time-invariant dynamic filters with fading memory. Our results supplement recent research showing that SCRs in the complex domain can approximate, to arbitrary precision, any unrestricted linear reservoir with a non-linear readout. We furthermore introduce a novel method to drastically reduce the number of SCR units, making such highly constrained architectures natural candidates for low-complexity hardware implementations. Our findings are supported by empirical studies on real-world time series datasets., Comment: 19 pages, 5 figures

Our objectives are to map the filamentary network around the Fornax-Eridanus Complex and probe the influence of the local environment on galaxy morphology. We employ the novel machine-learning tool, 1-DREAM (1-Dimensional, Recovery, Extraction, and Analysis of Manifolds) to detect and model filaments around the Fornax cluster. We then use the morphology-density relation of galaxies to examine the variation in the galaxies' morphology with respect to their distance from the central axis of the detected filaments. We detect 27 filaments that vary in length and galaxy-number density around the Fornax-Eridanus Complex. These filaments showcase a variety of environments; some filaments encompass groups/clusters, while others are only inhabited by galaxies in pristine filamentary environments. We also reveal a well-known structure -- the Fornax Wall, that passes through the Dorado group, Fornax cluster, and Eridanus supergroup. Regarding the morphology of galaxies, we find that early-type galaxies (ETGs) populate high-density filaments and high-density regions of the Fornax Wall. Furthermore, the fraction of ETGs decreases as the distance to the filament spine increases. Of the total galaxy population in filaments, ~7% are ETGs and ~24% are late-type galaxies (LTGs) located in pristine environments of filaments, while ~27% are ETGs and ~42% are LTGs in groups/clusters within filaments. This study reveals the Cosmic Web around the Fornax Cluster and asserts that filamentary environments are heterogeneous in nature. When investigating the role of the environment on galaxy morphology, it is essential to consider both, the local number-density and a galaxy's proximity to the filament spine. Within this framework, we ascribe the observed morphological segregation in the Fornax Wall to pre-processing of galaxies within groups embedded in it., Comment: Accepted for publication in A&A. 21 pages with 15 figures

Nowadays, neural network (NN) and deep learning (DL) techniques are widely adopted in many applications, including recommender systems. Given the sparse and stochastic nature of collaborative filtering (CF) data, recent works have critically analyzed the effective improvement of neural-based approaches compared to simpler and often transparent algorithms for recommendation. Previous results showed that NN and DL models can be outperformed by traditional algorithms in many tasks. Moreover, given the largely black-box nature of neural-based methods, interpretable results are not naturally obtained. Following on this debate, we first present a transparent probabilistic model that topologically organizes user and product latent classes based on the review information. In contrast to popular neural techniques for representation learning, we readily obtain a statistical, visualization-friendly tool that can be easily inspected to understand user and product characteristics from a textual-based perspective. Then, given the limitations of common embedding techniques, we investigate the possibility of using the estimated interpretable quantities as model input for a rating prediction task. To contribute to the recent debates, we evaluate our results in terms of both capacity for interpretability and predictive performances in comparison with popular text-based neural approaches. The results demonstrate that the proposed latent class representations can yield competitive predictive performances, compared to popular, but difficult-to-interpret approaches.

This work proposes a time series prediction method based on the kernel view of linear reservoirs. In particular, the time series motifs of the reservoir kernel are used as representational basis on which general readouts are constructed. We provide a geometric interpretation of our approach shedding light on how our approach is related to the core reservoir models and in what way the two approaches differ. Empirical experiments then compare predictive performances of our suggested model with those of recent state-of-art transformer based models, as well as the established recurrent network model - LSTM. The experiments are performed on both univariate and multivariate time series and with a variety of prediction horizons. Rather surprisingly we show that even when linear readout is employed, our method has the capacity to outperform transformer models on univariate time series and attain competitive results on multivariate benchmark datasets. We conclude that simple models with easily controllable capacity but capturing enough memory and subsequence structure can outperform potentially over-complicated deep learning models. This does not mean that reservoir motif based models are preferable to other more complex alternatives - rather, when introducing a new complex time series model one should employ as a sanity check simple, but potentially powerful alternatives/baselines such as reservoir models or the models introduced here., Comment: 8 pages

Growing attention has been brought to the fact that many real directed networks exhibit hierarchy and directionality as measured through techniques like Trophic Analysis and non-normality. We propose a simple growing network model where the probability of connecting to a node is defined by a preferential attachment mechanism based on degree and the difference in fitness between nodes. In particular, we show how mechanisms such as degree-based preferential attachment and node fitness interactions can lead to the emergence of the spectrum of hierarchy and directionality observed in real networks. In this work, we study various features of this model relating to network hierarchy, as measured by Trophic Analysis. This includes (I) how preferential attachment can lead to network hierarchy, (II) how scale-free degree distributions and network hierarchy can coexist, (III) the correlation between node fitness and trophic level, (IV) how the fitness parameters can predict trophic incoherence and how the trophic level difference distribution compares to the fitness difference distribution, (V) the relationship between trophic level and degree imbalance and the unique role of nodes at the ends of the fitness hierarchy and (VI) how fitness interactions and degree-based preferential attachment can interplay to generate networks of varying coherence and degree distribution. We also provide an example of the intuition this work enables in the analysis of a real historical network. This work provides insight into simple mechanisms which can give rise to hierarchy in directed networks and quantifies the usefulness and limitations of using Trophic Analysis as an analysis tool for real networks., Comment: 50 pages, 15 figures

The halo of the Milky Way galaxy hosts multiple dynamically coherent substructures known as stellar streams that are remnants of tidally disrupted systems such as globular clusters (GCs) and dwarf galaxies (DGs). A particular case is that of the Jhelum stream, which is known for its complex morphology. Using the available data from Gaia DR3, we extracted a region on the sky that contains Jhelum. We then applied the novel Locally Aligned Ant Technique (LAAT) on the position and proper motion space of stars belonging to the selected region to highlight the stars that are closely aligned with a local manifold in the data and the stars belonging to regions of high local density. We find that the overdensity representing the stream in proper motion space is composed of two components, and show the correspondence of these two signals to the previously reported narrow and broad spatial components of Jhelum. We made use of the radial velocity measurements provided by the $S^5$ survey to confirm, for the first time, a separation between the two components in radial velocity. We show that the narrow and broad components have velocity dispersions of $4.84^{+1.23}_{-0.79}$~km/s and $19.49^{+2.19}_{-1.84}$~km/s, and metallicity dispersions of $0.15^{+0.18}_{-0.10}$ and $0.34^{+0.13}_{-0.09}$, respectively. These measurements, and the difference in component widths, could be explained with a scenario where Jhelum is the remnant of a GC embedded within a DG that were accreted onto the Milky Way during their infall. Although the properties of Jhelum can be explained with this merger scenario, other progenitors of the narrow component remain possible such as a nuclear star cluster or a DG. To rule these possibilities out, we would need more observational data of member stars of the stream. Our analysis highlights the importance of the internal structure of streams with regards to their formation history.

Reservoir computation models form a subclass of recurrent neural networks with fixed non-trainable input and dynamic coupling weights. Only the static readout from the state space (reservoir) is trainable, thus avoiding the known problems with propagation of gradient information backwards through time. Reservoir models have been successfully applied in a variety of tasks and were shown to be universal approximators of time-invariant fading memory dynamic filters under various settings. Simple cycle reservoirs (SCR) have been suggested as severely restricted reservoir architecture, with equal weight ring connectivity of the reservoir units and input-to-reservoir weights of binary nature with the same absolute value. Such architectures are well suited for hardware implementations without performance degradation in many practical tasks. In this contribution, we rigorously study the expressive power of SCR in the complex domain and show that they are capable of universal approximation of any unrestricted linear reservoir system (with continuous readout) and hence any time-invariant fading memory filter over uniformly bounded input streams., Comment: 28 pages

The distribution of galaxies and clusters of galaxies on the mega-parsec scale of the Universe follows an intricate pattern now famously known as the Large-Scale Structure or the Cosmic Web. To study the environments of this network, several techniques have been developed that are able to describe its properties and the properties of groups of galaxies as a function of their environment. In this work we analyze the previously introduced framework: 1-Dimensional Recovery, Extraction, and Analysis of Manifolds (1-DREAM) on N-body cosmological simulation data of the Cosmic Web. The 1-DREAM toolbox consists of five Machine Learning methods, whose aim is the extraction and modelling of 1-dimensional structures in astronomical big data settings. We show that 1-DREAM can be used to extract structures of different density ranges within the Cosmic Web and to create probabilistic models of them. For demonstration, we construct a probabilistic model of an extracted filament and move through the structure to measure properties such as local density and velocity. We also compare our toolbox with a collection of methodologies which trace the Cosmic Web. We show that 1-DREAM is able to split the network into its various environments with results comparable to the state-of-the-art methodologies. A detailed comparison is then made with the public code DisPerSE, in which we find that 1-DREAM is robust against changes in sample size making it suitable for analyzing sparse observational data, and finding faint and diffuse manifolds in low density regions.

Knowing which nodes are influential in a complex network and whether the network can be influenced by a small subset of nodes is a key part of network analysis. However, many traditional measures of importance focus on node level information without considering the global network architecture. We use the method of Trophic Analysis to study directed networks and show that both "influence" and "influenceability" in directed networks depend on the hierarchical structure and the global directionality, as measured by the trophic levels and trophic coherence, respectively. We show that in directed networks trophic hierarchy can explain: the nodes that can reach the most others; where the eigenvector centrality localises; which nodes shape the behaviour in opinion or oscillator dynamics; and which strategies will be successful in generalised rock-paper-scissors games. We show, moreover, that these phenomena are mediated by the global directionality. We also highlight other structural properties of real networks related to influenceability, such as the pseudospectra, which depend on trophic coherence. These results apply to any directed network and the principles highlighted, that node hierarchy is essential for understanding network influence, mediated by global directionality, are applicable to many real-world dynamics., Comment: 31 pages, 28 figures

In many real, directed networks, the strongly connected component of nodes which are mutually reachable is very small. This does not fit with current theory, based on random graphs, according to which strong connectivity depends on mean degree and degree-degree correlations. And it has important implications for other properties of real networks and the dynamical behaviour of many complex systems. We find that strong connectivity depends crucially on the extent to which the network has an overall direction or hierarchical ordering -- a property measured by trophic coherence. Using percolation theory, we find the critical point separating weakly and strongly connected regimes, and confirm our results on many real-world networks, including ecological, neural, trade and social networks. We show that the connectivity structure can be disrupted with minimal effort by a targeted attack on edges which run counter to the overall direction. And we illustrate with example dynamics -- the SIS model, majority vote, Kuramoto oscillators and the voter model -- how a small number of edge deletions can utterly change dynamical processes in a wide range of systems., Comment: 16 pages, 6 figures

Application of interpretable machine learning techniques on medical datasets facilitate early and fast diagnoses, along with getting deeper insight into the data. Furthermore, the transparency of these models increase trust among application domain experts. Medical datasets face common issues such as heterogeneous measurements, imbalanced classes with limited sample size, and missing data, which hinder the straightforward application of machine learning techniques. In this paper we present a family of prototype-based (PB) interpretable models which are capable of handling these issues. The models introduced in this contribution show comparable or superior performance to alternative techniques applicable in such situations. However, unlike ensemble based models, which have to compromise on easy interpretation, the PB models here do not. Moreover we propose a strategy of harnessing the power of ensembles while maintaining the intrinsic interpretability of the PB models, by averaging the model parameter manifolds. All the models were evaluated on a synthetic (publicly available dataset) in addition to detailed analyses of two real-world medical datasets (one publicly available). Results indicated that the models and strategies we introduced addressed the challenges of real-world medical data, while remaining computationally inexpensive and transparent, as well as similar or superior in performance compared to their alternatives.

Many real-world networks are directed, sparse and hierarchical, with a mixture of feed-forward and feedback connections with respect to the hierarchy. Moreover, a small number of 'master' nodes are often able to drive the whole system. We study the dynamics of pattern presentation and recovery on sparse, directed, Hopfield-like neural networks using Trophic Analysis to characterise their hierarchical structure. This is a recent method which quantifies the local position of each node in a hierarchy (trophic level) as well as the global directionality of the network (trophic coherence). We show that even in a recurrent network, the state of the system can be controlled by a small subset of neurons which can be identified by their low trophic levels. We also find that performance at the pattern recovery task can be significantly improved by tuning the trophic coherence and other topological properties of the network. This may explain the relatively sparse and coherent structures observed in the animal brain, and provide insights for improving the architectures of artificial neural networks. Moreover, we expect that the principles we demonstrate, through numerical analysis, here will be relevant for a broad class of system whose underlying network structure is directed and sparse, such as biological, social or financial networks., Comment: 14 pages, 15 figures

In this paper, we develop a new classification method for manifold-valued data in the framework of probabilistic learning vector quantization. In many classification scenarios, the data can be naturally represented by symmetric positive definite matrices, which are inherently points that live on a curved Riemannian manifold. Due to the non-Euclidean geometry of Riemannian manifolds, traditional Euclidean machine learning algorithms yield poor results on such data. In this paper, we generalize the probabilistic learning vector quantization algorithm for data points living on the manifold of symmetric positive definite matrices equipped with Riemannian natural metric (affine-invariant metric). By exploiting the induced Riemannian distance, we derive the probabilistic learning Riemannian space quantization algorithm, obtaining the learning rule through Riemannian gradient descent. Empirical investigations on synthetic data, image data , and motor imagery EEG data demonstrate the superior performance of the proposed method., Comment: 15 pages, 7 figures

Along with the great success of deep neural networks, there is also growing concern about their black-box nature. The interpretability issue affects people's trust on deep learning systems. It is also related to many ethical problems, e.g., algorithmic discrimination. Moreover, interpretability is a desired property for deep networks to become powerful tools in other research fields, e.g., drug discovery and genomics. In this survey, we conduct a comprehensive review of the neural network interpretability research. We first clarify the definition of interpretability as it has been used in many different contexts. Then we elaborate on the importance of interpretability and propose a novel taxonomy organized along three dimensions: type of engagement (passive vs. active interpretation approaches), the type of explanation, and the focus (from local to global interpretability). This taxonomy provides a meaningful 3D view of distribution of papers from the relevant literature as two of the dimensions are not simply categorical but allow ordinal subcategories. Finally, we summarize the existing interpretability evaluation methods and suggest possible research directions inspired by our new taxonomy., Comment: This work has been accepted by IEEE-TETCI

This paper presents a geometric approach to pitch estimation (PE)-an important problem in Music Information Retrieval (MIR), and a precursor to a variety of other problems in the field. Though there exist a number of highly-accurate methods, both mono-pitch estimation and multi-pitch estimation (particularly with unspecified polyphonic timbre) prove computationally and conceptually challenging. A number of current techniques, whilst incredibly effective, are not targeted towards eliciting the underlying mathematical structures that underpin the complex musical patterns exhibited by acoustic musical signals. Tackling the approach from both a theoretical and experimental perspective, we present a novel framework, a basis for further work in the area, and results that (whilst not state of the art) demonstrate relative efficacy. The framework presented in this paper opens up a completely new way to tackle PE problems, and may have uses both in traditional analytical approaches, as well as in the emerging machine learning (ML) methods that currently dominate the literature.

Dimensionality reduction and clustering are often used as preliminary steps for many complex machine learning tasks. The presence of noise and outliers can deteriorate the performance of such preprocessing and therefore impair the subsequent analysis tremendously. In manifold learning, several studies indicate solutions for removing background noise or noise close to the structure when the density is substantially higher than that exhibited by the noise. However, in many applications, including astronomical datasets, the density varies alongside manifolds that are buried in a noisy background. We propose a novel method to extract manifolds in the presence of noise based on the idea of Ant colony optimization. In contrast to the existing random walk solutions, our technique captures points that are locally aligned with major directions of the manifold. Moreover, we empirically show that the biologically inspired formulation of ant pheromone reinforces this behavior enabling it to recover multiple manifolds embedded in extremely noisy data clouds. The algorithm performance in comparison to state-of-the-art approaches for noise reduction in manifold detection and clustering is demonstrated, on several synthetic and real datasets, including an N-body simulation of a cosmological volume., Comment: Accepted for publication by IEEE Transactions on Knowledge and Data Engineering

Increasing number of sectors which affect human lives, are using Machine Learning (ML) tools. Hence the need for understanding their working mechanism and evaluating their fairness in decision-making, are becoming paramount, ushering in the era of Explainable AI (XAI). In this contribution we introduced a few intrinsically interpretable models which are also capable of dealing with missing values, in addition to extracting knowledge from the dataset and about the problem. These models are also capable of visualisation of the classifier and decision boundaries: they are the angle based variants of Learning Vector Quantization. We have demonstrated the algorithms on a synthetic dataset and a real-world one (heart disease dataset from the UCI repository). The newly developed classifiers helped in investigating the complexities of the UCI dataset as a multiclass problem. The performance of the developed classifiers were comparable to those reported in literature for this dataset, with additional value of interpretability, when the dataset was treated as a binary class problem., Comment: Accepted for proceedings of the International Joint Conference on Neural Networks (IJCNN) 2020

Reservoir computing is a popular approach to design recurrent neural networks, due to its training simplicity and approximation performance. The recurrent part of these networks is not trained (e.g., via gradient descent), making them appealing for analytical studies by a large community of researchers with backgrounds spanning from dynamical systems to neuroscience. However, even in the simple linear case, the working principle of these networks is not fully understood and their design is usually driven by heuristics. A novel analysis of the dynamics of such networks is proposed, which allows the investigator to express the state evolution using the controllability matrix. Such a matrix encodes salient characteristics of the network dynamics; in particular, its rank represents an input-indepedent measure of the memory capacity of the network. Using the proposed approach, it is possible to compare different reservoir architectures and explain why a cyclic topology achieves favourable results as verified by practitioners.

Advances in machine learning technologies have led to increasingly powerful models in particular in the context of big data. Yet, many application scenarios demand for robustly interpretable models rather than optimum model accuracy; as an example, this is the case if potential biomarkers or causal factors should be discovered based on a set of given measurements. In this contribution, we focus on feature selection paradigms, which enable us to uncover relevant factors of a given regularity based on a sparse model. We focus on the important specific setting of linear ordinal regression, i.e.\ data have to be ranked into one of a finite number of ordered categories by a linear projection. Unlike previous work, we consider the case that features are potentially redundant, such that no unique minimum set of relevant features exists. We aim for an identification of all strongly and all weakly relevant features as well as their type of relevance (strong or weak); we achieve this goal by determining feature relevance bounds, which correspond to the minimum and maximum feature relevance, respectively, if searched over all equivalent models. In addition, we discuss how this setting enables us to substitute some of the features, e.g.\ due to their semantics, and how to extend the framework of feature relevance intervals to the setting of privileged information, i.e.\ potentially relevant information is available for training purposes only, but cannot be used for the prediction itself., Comment: Preprint accepted at Neurocomputing

We present Extended Riemannian Stochastic Derivative-Free Optimization (Extended RSDFO), a novel population-based stochastic optimization algorithm on Riemannian manifolds that addresses the locality and implicit assumptions of manifold optimization in the literature. We begin by investigating the Information Geometrical structure of statistical model over Riemannian manifolds. This establishes a geometrical framework of Extended RSDFO using both the statistical geometry of the decision space and the Riemannian geometry of the search space. We construct locally inherited probability distribution via an orientation-preserving diffeomorphic bundle morphism, and then extend the information geometrical structure to mixture densities over totally bounded subsets of manifolds. The former relates the information geometry of the decision space and the local point estimations on the search space manifold. The latter overcomes the locality of parametric probability distributions on Riemannian manifolds. We then construct Extended RSDFO and study its structure and properties from a geometrical perspective. We show that Extended RSDFO's expected fitness improves monotonically and it's global eventual convergence in finitely many steps on connected compact Riemannian manifolds. Extended RSDFO is compared to state-of-the-art manifold optimization algorithms on multi-modal optimization problems over a variety of manifolds. In particular, we perform a novel synthetic experiment on Jacob's ladder to motivate and necessitate manifold optimization. Jacob's ladder is a non-compact manifold of countably infinite genus, which cannot be expressed as polynomial constraints and does not have a global representation in an ambient Euclidean space. Optimization problems on Jacob's ladder thus cannot be addressed by traditional (constraint) optimization methods on Euclidean spaces., Comment: The present abstract is slightly altered from the PDF version due to the limitation "The abstract field cannot be longer than 1,920 characters"

Parameterized state space models in the form of recurrent networks are often used in machine learning to learn from data streams exhibiting temporal dependencies. To break the black box nature of such models it is important to understand the dynamical features of the input driving time series that are formed in the state space. We propose a framework for rigorous analysis of such state representations in vanishing memory state space models such as echo state networks (ESN). In particular, we consider the state space a temporal feature space and the readout mapping from the state space a kernel machine operating in that feature space. We show that: (1) The usual ESN strategy of randomly generating input-to-state, as well as state coupling leads to shallow memory time series representations, corresponding to cross-correlation operator with fast exponentially decaying coefficients; (2) Imposing symmetry on dynamic coupling yields a constrained dynamic kernel matching the input time series with straightforward exponentially decaying motifs or exponentially decaying motifs of the highest frequency; (3) Simple cycle high-dimensional reservoir topology specified only through two free parameters can implement deep memory dynamic kernels with a rich variety of matching motifs. We quantify richness of feature representations imposed by dynamic kernels and demonstrate that for dynamic kernel associated with cycle reservoir topology, the kernel richness undergoes a phase transition close to the edge of stability., Comment: 45 pages, 17 figures, accepted

Astronomical observations already produce vast amounts of data through a new generation of telescopes that cannot be analyzed manually. Next-generation telescopes such as the Large Synoptic Survey Telescope and the Square Kilometer Array are planned to become operational in this decade and the next, and will increase the data volume by many orders of magnitude. The increased spatial, temporal and spectral resolution afford a powerful magnifying lens on the physical processes that underlie the data but, at the same time, generate unprecedented complexity hard to exploit for knowledge extraction. It is therefore imperative to develop machine intelligence, machine learning (ML) in particular, suitable for processing the amount and variety of astronomical data that will be collected, and capable of answering scientific questions based on the data. Astronomical data exhibit the usual challenges associated with 'big data' such as immense volumes, high dimensionality, missing or highly distorted observations. In addition, astronomical data can exhibit large continuous observational gaps, very low signal-to-noise ratio and the need to distinguish between true missing data and non-detections due to upper limits). There are strict laws of physics behind the data production which can be assimilated into ML mechanisms to improve over general off-the-shelf state-of-the-art methods. Significant progress in the face of these challenges can be achieved only via the new discipline of Astroinformatics: a symbiosis of diverse disciplines, such as ML, probabilistic modeling, astronomy and astrophysics, statistics, distributed computing and natural computation. This editorial summarizes the contents of a soon to appear Focus Issue of the PASP on Machine Learning in Astronomy and Astrophysics (with contributions by 69 authors representing 15 countries, from 6 continents)., Comment: 11 pages

The mass and spin distributions of compact binary gravitational-wave sources are currently uncertain due to complicated astrophysics involved in their formation. Multiple sub-populations of compact binaries representing different evolutionary scenarios may be present among sources detected by Advanced LIGO and Advanced Virgo. In addition to hierarchical modelling, unmodelled methods can aid in determining the number of sub-populations and their properties. In this paper, we apply Gaussian mixture model clustering to 1000 simulated gravitational-wave compact binary sources from a mixture of five sub-populations. Using both mass and spin as input parameters, we determine how many binary detections are needed to accurately determine the number of sub-populations and their mass and spin distributions. In the most difficult case that we consider, where two sub-populations have identical mass distributions but differ in their spin, which is poorly constrained by gravitational-wave detections, we find that ~ 400 detections are needed before we can identify the correct number of sub-populations.

Weather and atmospheric patterns are often persistent. The simplest weather forecasting method is the so-called persistence model, which assumes that the future state of a system will be similar (or equal) to the present state. Machine learning (ML) models are widely used in different weather forecasting applications, but they need to be compared to the persistence model to analyse whether they provide a competitive solution to the problem at hand. In this paper, we devise a new model for predicting low-visibility in airports using the concepts of mixture of experts. Visibility level is coded as two different ordered categorical variables: cloud height and runway visual height. The underlying system in this application is stagnant approximately in 90% of the cases, and standard ML models fail to improve on the performance of the persistence model. Because of this, instead of trying to simply beat the persistence model using ML, we use this persistence as a baseline and learn an ordinal neural network model that refines its results by focusing on learning weather fluctuations. The results show that the proposal outperforms persistence and other ordinal autoregressive models, especially for longer time horizon predictions and for the runway visual height variable., Comment: Published in IEEE International Joint Conference on Neural Networks (IJCNN) 2018

Data augmentation is rapidly gaining attention in machine learning. Synthetic data can be generated by simple transformations or through the data distribution. In the latter case, the main challenge is to estimate the label associated to new synthetic patterns. This paper studies the effect of generating synthetic data by convex combination of patterns and the use of these as unsupervised information in a semi-supervised learning framework with support vector machines, avoiding thus the need to label synthetic examples. We perform experiments on a total of 53 binary classification datasets. Our results show that this type of data over-sampling supports the well-known cluster assumption in semi-supervised learning, showing outstanding results for small high-dimensional datasets and imbalanced learning problems., Comment: Published in the Thirty-Third AAAI Conference on Artificial Intelligence, 2019

The increasing occurrence of ordinal data, mainly sociodemographic, led to a renewed research interest in ordinal regression, i.e. the prediction of ordered classes. Besides model accuracy, the interpretation of these models itself is of high relevance, and existing approaches therefore enforce e.g. model sparsity. For high dimensional or highly correlated data, however, this might be misleading due to strong variable dependencies. In this contribution, we aim for an identification of feature relevance bounds which - besides identifying all relevant features - explicitly differentiates between strongly and weakly relevant features., Comment: preprint of a paper accepted for oral presentation at the 27th European Symposium on Artificial Neural Networks (ESANN 2019)

Predicting investors reactions to financial and political news is important for the early detection of stock market jitters. Evidence from several recent studies suggests that online social media could improve prediction of stock market movements. However, utilizing such information to predict strong stock market fluctuations has not been explored so far. In this work, we propose a novel event detection method on Twitter, tailored to detect financial and political events that influence a specific stock market. The proposed approach applies a bursty topic detection method on a stream of tweets related to finance or politics followed by a classification process which filters-out events that do not influence the examined stock market. We train our classifier to recognise real events by using solely information about stock market volatility, without the need of manual labeling. We model Twitter events as feature vectors that encompass a rich variety of information, such as the geographical distribution of tweets, their polarity, information about their authors as well as information about bursty words associated with the event. We show that utilizing only information about tweets polarity, like most previous studies, results in wasting important information. We apply the proposed method on high-frequency intra-day data from the Greek and Spanish stock market and we show that our financial event detector successfully predicts most of the stock market jitters.

The abundance of data produced daily from large variety of sources has boosted the need of novel approaches on causal inference analysis from observational data. Observational data often contain noisy or missing entries. Moreover, causal inference studies may require unobserved high-level information which needs to be inferred from other observed attributes. In such cases, inaccuracies of the applied inference methods will result in noisy outputs. In this study, we propose a novel approach for causal inference when one or more key variables are noisy. Our method utilizes the knowledge about the uncertainty of the real values of key variables in order to reduce the bias induced by noisy measurements. We evaluate our approach in comparison with existing methods both on simulated and real scenarios and we demonstrate that our method reduces the bias and avoids false causal inference conclusions in most cases., Comment: In Proceedings of International Joint Conference Of Neural Networks (IJCNN) 2017