Author name cluster

Inés Couso

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

6 papers

2 author rows

EAAI Journal 2025 Journal Article

Data imputation in the frequency domain using Echo State Networks

Luciano Sánchez
Nahuel Costa
Inés Couso

This study addresses the problem of reconstructing time series data with only partial spectral information, specifically the power spectral density (PSD), without corresponding phase details. Such situations are common in industrial settings where sensors measure the energy across various frequency bands but do not record the temporal signal. Existing phase-imputation methods often fail to produce reliable time-domain signals suitable for system identification algorithms, such as those used for calculating vibration modes and structural health monitoring. To address this issue, a solution is proposed using a physics-informed Echo State Network (ESN) designed to impute a time signal whose PSD matches the observed spectral data. The approach uses fuzzy sets to represent our limited knowledge about the inputs and extends the ESN definition to handle fuzzy-valued inputs, integrating physical insights directly into the frequency domain within the loss function. This enables the simultaneous learning of a set of weak constraints on the amplitude and autocorrelation of the unknown excitation to the physical system, alongside a recurrent neural network that models the functional dependence between these inputs and the partially observed outputs. This methodology is validated through two empirical analyses: one using a synthetic dataset designed to simulate real-world scenarios for benchmarking, and another through a practical case study focused on diagnosing the condition of industrial fans using vibration data. Empirical validation demonstrates that the proposed method successfully identified shifts in vibration modes, ranging from 8% to 17%, following structural changes.

Details DOI

EAAI Journal 2017 Journal Article

A class of Monotone Fuzzy rule-based Wiener systems with an application to Li-ion battery modelling

Luciano Sánchez
Inés Couso
Cecilio Blanco

A class of Fuzzy rule-based Monotone Wiener Models (FMWMs) is introduced. These are transformation models comprising a linear dynamical block and a memoryless nonlinearity. The smoothest dynamical block that has an output which is comonotonic with the training data is sought. The dependence between the output of the linear block and the output of the system is described via a set of fuzzy rules. This paper considers systems with a sensitive dependence on the initial conditions and also with a moderate amount of uncertainty in the initial state. A new learning algorithm is proposed that makes use of recent statistical tests for assessing the comonotonicity of imprecisely perceived sequences of data. The main aim of the proposed models is to estimate different health parameters of rechargeable batteries for automotive use. For this practical application, FMWMs are shown to improve a selection of models with a varying degree of embedded domain knowledge, ranging from first-principles models to universal approximators.

Details DOI

ISIPTA Conference 2017 Conference Paper

Maximum Likelihood with Coarse Data based on Robust Optimisation

Romain Guillaume
Inés Couso
Didier Dubois

This paper deals with the problem of probability estimation in the context of coarse data. Probabilities are estimated using the maximum likelihood principle. Our approach presupposes that each imprecise observation underlies a precise one, and that the uncertainty that pervades its observation is epistemic, rather than representing noise. As a consequence, the likelihood function of the ill-observed sample is set-valued. In this paper, we apply a robust optimization method to find a safe plausible estimate of the probabilities of elementary events on finite state spaces. More precisely we use a maximin criterion on the imprecise likelihood function. We show that there is a close connection between the robust maximum likelihood strategy and the maximization of entropy among empirical distributions compatible with the incomplete data. A mathematical model in terms of maximal flow on graphs, based on duality theory, is proposed. It results in a linear objective function and convex constraints. This result is somewhat surprizing since maximum entropy problems are known to be complex due to the maximization of a concave function on a convex set.

PDF Details

ISIPTA Conference 2017 Conference Paper

Reconciling Bayesian and Frequentist Tests: the Imprecise Counterpart

Inés Couso
Antonio Álvarez-Caballero
Luciano Sánchez

Imprecise Dirichlet Process-based tests (IDP-tests, for short) have been recently introduced in the literature. They overcome the problem of deciding how to select a single prior in Bayesian hypothesis testing, in the absence of prior information. They make use of a “near-ignorance” model, that behaves a priori as a vacuous model for some basic inferences, but it provides non-vacuous posterior inferences. We perform empirical studies regarding the behavior of IDP-tests for the particular case of Wilcoxon rank sum test. We show that the upper and lower posterior probabilities can be expressed as tail probabilities based on the value of the $U$ statistic. We construct an imprecise frequentist-based test that reproduces the same decision rule as the the IDP test. It considers a neighbourhood around the $U$-statistic value. If all the values in the neighbourhood belong to the rejection zone (resp. to the acceptance region), the null hypothesis is rejected (resp. accepted). Otherwise, the judgement is suspended.

PDF Details

ICML Conference 2017 Conference Paper

Statistical Inference for Incomplete Ranking Data: The Case of Rank-Dependent Coarsening

Mohsen Ahmadi Fahandar
Eyke Hüllermeier
Inés Couso

We consider the problem of statistical inference for ranking data, specifically rank aggregation, under the assumption that samples are incomplete in the sense of not comprising all choice alternatives. In contrast to most existing methods, we explicitly model the process of turning a full ranking into an incomplete one, which we call the coarsening process. To this end, we propose the concept of rank-dependent coarsening, which assumes that incomplete rankings are produced by projecting a full ranking to a random subset of ranks. For a concrete instantiation of our model, in which full rankings are drawn from a Plackett-Luce distribution and observations take the form of pairwise preferences, we study the performance of various rank aggregation methods. In addition to predictive accuracy in the finite sample setting, we address the theoretical question of consistency, by which we mean the ability to recover a target ranking when the sample size goes to infinity, despite a potential bias in the observations caused by the (unknown) coarsening.

Details

EAAI Journal 2015 Journal Article

Sequential pattern mining applied to aeroengine condition monitoring with uncertain health data

Ana Palacios
Alvaro Martínez
Luciano Sánchez
Inés Couso

Numerical algorithms that can assess Engine Health Monitoring (EHM) data in aeroengines are influenced by the high level of uncertainty inherent to gas path measurements and engine-to-engine variability. Among them, fuzzy rule-based techniques have been successfully used due to their robustness towards noisy signals and their capability to learn human-readable rules from data. These techniques are useful in detecting the presence of certain types of abnormal events or general engine deterioration, through the identification of specific combinations of EHM signals associated with these specific cases. However, there are also other types of engine events that manifest themselves as an ordered sequence of otherwise normal combinations of the EHM signals. These combinations are dismissed when considered in isolation as the current existing techniques cannot assess them. In this paper it is proposed to use sequence mining techniques in order to obtain fuzzy rules from uncertain EHM data which can in turn be used to identify the cases where an engine event is determined as a sequence of otherwise normal combinations of EHM signals. The results are subsequently tested on a representative sample of aeroengine data.

Details DOI