In many real-world applications, data are represented by multiple instances and simultaneously associated with multiple labels. These data are always redundant and generally contaminated by different noise levels. As a result, several machine learning models fail to achieve good classification and find an optimal mapping. Feature selection, instance selection, and label selection are three effective dimensionality reduction techniques. Nevertheless, the literature was limited to feature and/or instance selection but has, to some extent, neglected label selection, which also plays an essential role in the preprocessing step, as label noises can adversely affect the performance of the underlying learning algorithms. In this article, we propose a novel framework termed multilabel Feature Instance Label Selection (mFILS) that simultaneously performs feature, instance, and label selections in both convex and nonconvex scenarios. To the best of our knowledge, this article offers, for the first time ever, a study using the triple and simultaneous selection of features, instances, and labels based on convex and nonconvex penalties in a multilabel scenario. Experimental results are built on some known benchmark datasets to validate the effectiveness of the proposed mFILS.

Feature selection is a very important part of successful data mining applications, because it contributes to the simplification and explainability of the resulting model through dimensionality reduction. Feature selection methods exist for semi-supervised scenarios, in which a small subset of labeled instances is combined with a large batch of unlabeled instances to improve the learning performance. In a multi-label setting, feature selection identifies important features for all labels at once. However, some labels can also be irrelevant and then deteriorate feature selection. In a robust learning framework, removing these noisy labels can decrease the generalization error for the resulting regression model. In this work, we present a novel algorithm performing multi-label semi-supervised feature selection and label seleciton, and show its effectiveness on some publicly available datasets.

Assessing the quality of 360-degree images based on individual regions presents a challenging task. The lack of ground truth opinion scores (MOS) for specific regions makes it difficult to evaluate image quality accurately. Existing datasets only provide MOS for entire 360-degree images, which limits the granularity of assessment. To overcome this challenge, we propose a novel framework that employs adaptive patch labeling techniques. We leverage a set of 2D-IQA methods to generate quality score distributions for each patch in the 360-degree images. These distributions, combined with the available MOS, serve as labels for individual patches, providing a more comprehensive characterization of patch quality. Furthermore, we use these labels to adaptively select and refine deep neural features. By selectively choosing label-specific features, we enhance the accuracy and effectiveness of patch-based 360-degree image quality assessment. This approach allows us to focus on the most relevant and informative features for each patch, resulting in improved assessment performance. The experimental results on two benchmark datasets demonstrate that adaptive patch labeling and feature selection achieve accurate and reliable performances, thus advancing the field of 360-degree image quality assessment.

Multivariate time series data are invasive in different domains, ranging from data center supervision and e-commerce data to financial transactions. This kind of data presents an important challenge for anomaly detection due to the temporal dependency aspect of its observations. In this article, we investigate the problem of unsupervised local anomaly detection in multivariate time series data from temporal modeling and residual analysis perspectives. The residual analysis has been shown to be effective in classical anomaly detection problems. However, it is a nontrivial task in multivariate time series as the temporal dependency between the time series observations complicates the residual modeling process. Methodologically, we propose a unified learning framework to characterize the residuals and their coherence with the temporal aspect of the whole multivariate time series. Experiments on real-world datasets are provided showing the effectiveness of the proposed algorithm.

Traffic congestion is a phenomenon with which most drivers are familiar. It negatively affects drivers, city's residents and also the economic efficiency. Road pavement condition is one of the main causes of traffic congestion. Warning drivers of road pavement conditions is an effective and highly recommended solution to prevent a traffic jam before it actually occurs. Through this paper, we present a new strategy to reduce traffic congestion. We propose an unsupervised anomaly detection framework that predicts road obstacles. We suggest a well-motivated combination between the Long Short Term Memory (LSTM) and Local outlier value factor (LOF) techniques for predicting obstacles in time series. This combination can potentially offer very promising results in terms of prediction accuracy. We illustrate significant performance gains achieved by our technique with respect to the conventional methods.

In this paper, we aim to improve the performance, time complexity and energy efficiency of deep convolutional neural networks (CNNs) by combining hardware and specialization techniques. Since the pooling step represents a process that contributes significantly to CNNs performance improvement, we propose the Mode-Fisher (MF) pooling method. This form of pooling can potentially offer a very promising results in terms of improving feature extraction performance. The proposed method reduces significantly the data movement in the CNN and save up to 10% of total energy, without any performance penalty.

In this paper, we present an improved approach for semi-supervised regression problems. Our proposal is based on both, the use of the top eigen functions of integral operator derived from both labeled and unlabeled examples as the basis functions; and the learning of the prediction function by a Laplacian regularized regression. We compare our method with some representative ones dealing with semi-supervised regression. This comparison is done over several public data sets. We also verify the effectiveness of the proposed algorithm to reconstitute the installation date of the pipes of the Lyon Metropolis sewer network.

Temporal data presents challenges and opportunities for machine learning and data mining communities. Recently, the unsupervised anomaly detection task for this kind of data has received much attention. In this paper, we propose a new embedded approach, named ℓ2-DAT, for dealing with this task. Our approach consists in reframing this task as weighting-instance clustering problem based on Ridge regularization and Dynamic Time Warpping for time series data. The anomalous series are then detected by an optimization problem of a new proposed cost function. Extensive experiments on benchmark data sets are carried out for validating our approach and comparing it with other methods of detection.