Auswahl der wissenschaftlichen Literatur zum Thema „Soft-DTW“

Soft biometrics are already widely used as a support tool for user identification. However, it is not the only use for biometric information that is conceivable because such information can be sufficient to obtain minimal details from the user that are unrelated to his identity. Examples of what might be referred to as soft biometrics include gender, hand orientation, and emotional state. Utilizing physiologic modalities for soft-biometric work is extremely prevalent, prediction, but behavioral data is frequently disregarded. Keystroke dynamics and handwriting signature are two potential behavioral modalities that could be used to predict a user's gender, but they are rarely discussed in the literature together. This study seeks to fill this gap by examining the influence of combining these two distinct biometric modalities on the accuracy of gender prediction and the best way. Key Words: Item key-strokes, Bio-metric signatures, digital signs, dynamic temporal wrapping (DTW)

Obesity and dysphagia are of potential and direct serious harm to the human body health. A commonly used method is controlling food intake to avoid obesity or determining if dysphagia exists by monitoring the swallow . This paper proposes a swallow detecting principle based on nanometer materials sensor, and implements a wearable detecting system with advantage of improved DTW algorithm. The system efficiently detects and faithfully identifies swallowing. In addition, it reduces the demand for hardware computing power. The system meets the features of a wearable system, such as soft and comfortable, lightweight, portable, and noninvasive.

The current lack of harmonized standard test conditions for marine shipping hinders the comparison of performance and compliance assessments for different types of ships. This article puts forward a method for extracting ship loading cycles using machine learning algorithms. Time-series data are extracted from real ships in operation, and a segmented linear approximation method and a data normalization technique are adopted. A hierarchical-clustering type of soft dynamic time-warping similarity analysis method is presented to efficiently analyze the similarity of different time-series data, using soft dynamic time warping (Soft-DTW) combined with hierarchical clustering algorithms from the field of machine learning. The problem of data bias caused by spatial and temporal offset characteristics is effectively solved in marine test condition data. The validity and reliability of the proposed method are validated through the analysis of case data. The results demonstrate that the hierarchically clustered soft dynamic time-warping similarity analysis method can be considered reliable for obtaining test cases with different characteristics. Furthermore, it provides input conditions for effectively identifying the operating conditions of different types of ships with high levels of energy consumption and high emissions, thus allowing for the establishment of energy-saving and emissions-reducing sailing strategies.

Aiming to improve on the low accuracy of current transmission line icing prediction models and ignoring the objective law of icing of transmission lines, a transmission line icing prediction model considering the effect of transmission line tension on the bundle of icing thickness is proposed, based on a convolutional neural network (CNN) and bidirectional gated recurrent unit (BiGRU). Firstly, the finite element calculation model of the conductor and insulator system was established, and the change rule between transmission line tension and icing thickness was studied. Then, the convolutional neural network and bidirectional gated recurrent unit were used to construct a transmission line icing thickness prediction model The model incorporated a weighted fusion of soft−dynamic time warping (Soft−DTW) and the icing change rule as the loss function. Optimal weights were determined through the utilization of the grid search algorithm and cross−validation, contributing to an enhancement of the model’s generalization capabilities and a reduction in prediction errors. The results indicate that the proposed prediction model can consider the impact of line operating parameters, avoiding the shortcomings of prediction results conflicting with actual physical laws. Compared with traditional non−mechanical models, the proposed model showed reductions in root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) by 0.26–0.51%, 0.24–0.44%, and 5.77–13.33%, respectively, while the coefficient of determination (R2) increased by 0.07–0.13.

The accurate identification of the horizontal well marker bed is to guarantee the soft landing of the well trajectory. With the intelligent development of the petroleum industry, it is feasible to apply computers to identify the marker bed automatically. In case-based reasoning technology, the data of well logging while drilling (LWD) as characteristic parameters are compared with those of adjacent well. By taking the depth sequence of LWD data as time series and using Dynamic Time Warping (DTW) similarity measure algorithm, the similarity index of each drilling depth is calculated corresponding to the marker bed in the adjacent well. The total similarity curve is obtained by giving different weights of different feature parameters. Selecting natural gamma, deep resistivity, and shallow resistivity LWD curves as characteristic parameters, two horizontal wells in JL block of Junggar basin are analysed by this method. The result of similarity curve indicates the location of the marker bed and the total similarity value reaches 78%. The research shows that the method based on case-based reasoning can identify the marker bed of the horizontal well accurately and effectively, assist the geologist to carry out formation correlation of multiple wells at the same time, reduce the cost of human labour force, and improve work efficiency.

Accurately predicting the trajectories of mesoscale eddies is essential for comprehending the distribution of marine resources and the multiscale energy cascade in the ocean. Nevertheless, current approaches for predicting mesoscale eddy trajectories frequently exhibit inadequate examination of the intrinsic multiscale temporal data, resulting in diminished predictive precision. To address this challenge, our research introduces an enhanced transformer-based framework for predicting mesoscale eddy trajectories. Initially, a multivariate dataset of mesoscale eddy trajectories is constructed and expanded, encompassing eddy properties and pertinent ocean environmental information. Additionally, novel feature factors are delineated based on the physical attributes of eddies. Subsequently, a multi-head attention mechanism is introduced to bolster the modeling of the multiscale time-varying connections within eddy trajectories. Furthermore, the original positional encoding is substituted with Time-Absolute Position Encoding, which considers the dimensions and durations of the sequence mapping, thereby improving the distinguishability of embedded vectors. Ultimately, the Soft-DTW loss function is integrated to more accurately assess the overall discrepancies among mesoscale eddy trajectories, thereby improving the model’s resilience to erratic and diverse trajectory sequences. The effectiveness of the proposed framework is assessed using the eddy-abundant South China Sea. Our framework exhibits exceptional predictive accuracy, achieving a minimum central error of 8.507 km over a seven-day period, surpassing existing state-of-the-art models.

Abstract Background CD patients exhibit highly variable responses to biologics. While some patients achieve sustained remission with only one biologic over the course of their disease, other will require the sequencing of multiple biologics (“difficult-to-treat” patients) for optimal disease control. The aim of this study was to delineate the profile of treatment regimen of the biological exposed CD patients with luminal disease. Methods Among CD patients diagnosed between 1999 and 2019 with B1 phenotype at diagnosis, 203 patients who maintained this phenotype at maximum follow-up (FU, median 11y [8–15]) were selected, based on inclusion criteria. All events were recorded from date of diagnosis to maximum FU, from disease characteristics to treatment adaptations. A temporal clustering approach using k-means was applied to determine biologics treatment regimen profiles. Patient profiles were compared using the Dynamic Time Warping (DTW) similarity measure, considering 3 successive cumulative biologic exposures (for primary/secondary non response only). Clusters number was chosen to balance maximizing silhouette score while ensuring sufficient individuals per cluster. Cumulative biologic exposures were summarized as cluster profiles, represented by their corresponding barycenters computed with soft-DTW distance. Clustering approach was performed using Python (tslearn). Results The 203 patients were clustered in 5 distinct profiles, based on their cumulative biologics exposure in the first 10 years after CD diagnosis (Figure 1). Patients in Cluster 1 were not exposed to biologics (n=55). Patients in Cluster 2 were treated late with one biologic (n=51) while patients in Cluster 3 were treated early with one biologic (n=53). Patients from Cluster 4 (n=23) and Cluster 5 (n=21) were both treated early with multiple biologics. Patients from Cluster 2 were exposed to biologics later after CD diagnosis than patients from Cluster 3 to 5 (Median 5.7y [3.6–6.9] vs 0.7y [0.2-1.8], p<0.0001), as also shown by their lower biological exposure compared to those of Cluster 3 to 5 (median 48% [37-58] vs 80% [60-89], p<0.0001). Most interestingly, patients from Cluster 5 required faster biologics changes compared to patients from Cluster 4 (Figure 2). In Cluster 5 (rapid switchers), patients were exposed earlier compared to patients in Cluster 4 (Slow Switchers) to the first (p<0.005), second (P<0.005) and third biologic (p=0.01), introducing the concept of rapid and slow biologic switchers among CD patients. Conclusion This clustering approach highlights the highly variable pattern of response to biologics in luminal CD patients. Furthermore, this study discriminatessingle and multiple biologics-exposed patients, in whom we identified slow and rapid biologic switchers.

Short-term residential load forecasting plays a crucial role in smart grids, ensuring an optimal match between energy demands and generation. With the inherent volatility of residential load patterns, deep learning has gained attention due to its ability to capture complex nonlinear relationships within hidden layers. However, most existing studies have relied on default loss functions such as mean squared error (MSE) or mean absolute error (MAE) for neural networks. These loss functions, while effective in overall prediction accuracy, lack specialized focus on accurately predicting load peaks. This article presents a comparative analysis of soft-DTW loss function, a smoothed formulation of Dynamic Time Wrapping (DTW), compared to other commonly used loss functions, in order to assess its effectiveness in improving peak prediction accuracy. To evaluate peak performance, we introduce a novel evaluation methodology using confusion matrix and propose new errors for peak position and peak load, tailored specifically for assessing peak performance in short-term load forecasting. Our results demonstrate the superiority of soft-DTW in capturing and predicting load peaks, surpassing other commonly used loss functions. Furthermore, the combination of soft-DTW with other loss functions, such as soft-DTW + MSE, soft-DTW + MAE, and soft-DTW + TDI (Time Distortion Index), also enhances peak prediction. However, the differences between these combined soft-DTW loss functions are not substantial. These findings highlight the significance of utilizing specialized loss functions, like soft-DTW, to improve peak prediction accuracy in short-term load forecasting.

As a time-domain EEG feature reflecting the semantic processing of the human brain, the N400 event-related potentials still lack a mature classification and recognition scheme. To address the problems of low signal-to-noise ratio and difficult feature extraction of N400 data, we propose a Soft-DTW-based single-subject short-distance event-related potential averaging method by using the advantages of differentiable and efficient Soft-DTW loss function, and perform partial Soft-DTW averaging based on DTW distance within a single-subject range, and propose a Transformer-based ERP recognition classification model, which captures contextual information by introducing location coding and a self-attentive mechanism, combined with a Softmax classifier to classify N400 data. The experimental results show that the highest recognition accuracy of 0.8992 is achieved on the ERP-CORE N400 public dataset, verifying the effectiveness of the model and the averaging method.

Cette thèse aborde le problème de la détection d’anomalie dans les séries temporelles, en se focalisant sur les systèmes à comportement répétitif, tels que les robots industriels opérant sur des chaînes de production. La recherche traite plusieurs défis, notamment la quantité importante de données manquantes dans les jeux de données, ce qui entraîne un échantillonnage irrégulier des séries temporelles issues des capteurs, ainsi que des variations dans la durée de chaque répétition de la tâche.L'approche de détection d’anomalie présentée se déroule en trois étapes :- La première étape identifie les cycles répétitifs dans les séries temporelles entières et les segmente en sous-séquences correspondant à un cycle de la tâche malgré les éventuelles distorsions temporelles.- La deuxième étape calcule un prototype des cycles à l'aide d'un algorithme de barycentre optimisé par GPU spécifiquement adapté aux très grandes séries temporelles.- La troisième étape utilise le prototype pour détecter les cycles anormaux en calculant un score d'anomalie pour chaque cycle.L'approche globale, nommée WarpEd Time Series ANomaly Detection (WETSAND), utilise l'algorithme Dynamic Time Warping (Déformation Temporelle Dynamique) et ses variantes qui gèrent la nature déformée des séries temporelles.Les expériences ont été menées sur des robots manipulateurs réels des usines de Vitesco Technologies. Les robots manipulateurs représentent une part importante de l'automatisation dans l'industrie actuelle. Conçus pour effectuer des tâches spécifiques et répétitives en toute sécurité aux côtés des opérateurs humains, il est essentiel de prédire et de diagnostiquer toute déviation par rapport à leur comportement attendu. Par conséquent, surveiller le comportement de ces robots est crucial, car cela minimise les temps d'arrêt des chaînes de production et prolonge la durée de vie des systèmes en permettant d’ajuster les calendriers de maintenance. Dans l'ère numérique de l'Industrie 4.0, où la collecte, le stockage et le traitement des données sont omniprésents, les paramètres de ces robots sont surveillés en temps réel, garantissant ainsi l'exécution parfaite de leurs tâches.Les expériences montrent que WETSAND s'adapte à des signaux de grande taille, calcule des prototypes faciles à interpréter, fonctionne avec très peu de données et surpasse certaines approches de détection d'anomalie neuronales reconnues telles que les autoencodeurs. Une interface utilisateur basée sur le cloud a été conçue pour déployer WETSAND* dans les usines de Vitesco Technologies, où elle permet de surveiller en ligne différents robots sur les chaînes de production.Cette thèse fait partie du programme **CIFRE** sous la chaire « AI Collaborative: Transformations Synergiques en in diagnostic basé sur des modèles et sur des » au sein de l’institut interdisciplinaire d’intelligence artificielle ANITI (Artificial and Natural Intelligence Toulouse Institute). La recherche a été menée en collaboration entre le Laboratoire d'Analyse et d'Architecture des Systèmes (LAAS-CNRS) et Vitesco Technologies, situés à Toulouse, France
This thesis addresses the problem of detecting time series outliers, focusing on systems with repetitive behavior, such as industrial robots operating on production lines. The research addresses several challenges, notably the significant amount of missing data within the collected datasets that results in irregular sampling of the time series reported by sensors, as well as variations in the duration of each task repetition across the time series.The anomaly detection approach presented in this paper consists of three stages.- The first stage identifies the repetitive cycles in the lengthy time series and segments them into individual time series corresponding to one task cycle, while accounting for possible temporal distortions.- The second stage computes a prototype for the cycles using a GPU-based barycenter algorithm, specifically tailored for very large time series.- The third stage uses the prototype to detect abnormal cycles by computing an anomaly score for each cycle.The overall approach, named WarpEd Time Series ANomaly Detection (WETSAND), makes use of the Dynamic Time Warping algorithm and its variants because they are suited to the distorted nature of the time series.The experiments have been carried out with real robot manipulators of Vitesco Technology plants. Robot manipulators constitute a significant portion of automation in today’s industry. Designed to perform specific, repetitive tasks safely alongside human operators, it is essential to predict and diagnose any deviation from their expected behavior. Consequently, monitoring these robots' behavior is crucial, as it minimizes production line downtime and prolongs the system's lifespan through maintenance schedule adjustments. In the digital era of Industry 4.0, where data collection, storage, and processing are ubiquitous, the parameters of these robots are continuously monitored in real-time, ensuring their tasks are executed flawlessly.The experiments show that WETSAND scales to large signals, computes human-friendly prototypes, works with very little data, and outperforms some recognized neural anomaly detection approaches such as autoencoders. A cloud-based user interface has been designed to deploy WETSAND in the Vitesco Technologies plants and it monitors online different robots in the production chains.This thesis is part of CIFRE program under the “Collaborative AI : Synergistic transformations in model based and data-based diagnosis” chair at ANITI. The research has been conducted through a collaboration between the Laboratory of Analysis and Architecture of Systems (LAAS) and Vitesco Technologies, situated in Toulouse, France