Abstract: A method classifies missing labels. The method computes, using a neural network model trained on training data, rank-based statistics of a feature of a time series segment to attempt to select two candidate labels from the training data that the segment most likely belongs to. The method classifies the segment using k-NN-based classification applied to the training data, responsive to the two candidate labels being present in the training data. The method classifies the segment by hypothesis testing, responsive to only one candidate label being present in the training data. The method classifies the segment into a class with higher values of the rank-based statistics from among a plurality of classes with different values of the rank-based statistics, responsive to no candidate labels being present in the training data. The method corrects a prediction by an applicable one of the classifying steps by majority voting with time windows.

Abstract: Methods and systems for training a neural network include collecting model exemplar information from edge devices, each model exemplar having been trained using information local to the respective edge devices. The collected model exemplar information is aggregated together using federated averaging. Global model exemplars are trained using federated constrained clustering. The trained global exemplars are transmitted to respective edge devices.

Abstract: Systems and methods for retrieving similar multivariate time series segments are provided. The systems and methods include extracting a long feature vector and a short feature vector from a time series segment, converting the long feature vector into a long binary code, and converting the short feature vector into a short binary code. The systems and methods further include obtaining a subset of long binary codes from a binary dictionary storing dictionary long codes based on the short binary codes, and calculating similarity measure for each pair of the long feature vector with each dictionary long code. The systems and methods further include identifying a predetermined number of dictionary long codes having the similarity measures indicting a closest relationship between the long binary codes and dictionary long codes, and retrieving a predetermined number of time series segments associated with the predetermined number of dictionary long codes.

Abstract: Methods and systems of training and using a neural network model include training a time series embedding model and a text embedding model with unsupervised clustering to translate time series and text, respectively, to a shared latent space. The time series embedding model and the text embedding model are further trained using semi-supervised clustering that samples training data pairs of time series information and associated text for annotation.

Abstract: Methods and systems for detecting and correcting anomalies includes generating historical binary codes from historical time series segments. The historical time series segments are each made up of measurements from respective sensors. A latest binary code is generated from a latest time series segment. It is determined that the latest time series segment represents anomalous behavior, based on a comparison of the latest binary code to the historical binary codes. The sensors are ranked, based on a comparison of time series data of the sensors in the latest time series segment to respective time series data of the historical time series, to generate a sensor ranking. A corrective action is performed responsive to the detected anomaly, prioritized according to the sensor ranking.

Abstract: A first anomaly detection unit detects anomalous first monitored data from among a plurality of first monitored data obtained from a monitored system. A second anomaly detection unit operates in parallel with the first anomaly detection unit and detects anomalous second monitored data from among a plurality of second monitored data obtained from the monitored system. In a first storage unit, the anomalous first monitored data and the anomalous second monitored data detected before lapse of a given time from detection time of the anomalous first monitored data are stored in association with each other. A first determination unit, when the anomalous first monitored data is detected, retrieves the anomalous second monitored data associated with the detected anomalous first monitored data from the first storage unit and outputs a first anomaly detection result including the retrieved anomalous second monitored data and the detected anomalous first monitored data.

Abstract: Change points of a system represented by a plurality of time series are detected more appropriately. An information processing system includes means for learning, with respect to each of a plurality of time series, models that approximate partial time series respectively and are defined by parameters of the partial time series respectively, the partial time series being obtained by dividing a corresponding time series into a plurality of segments at change point candidates; and means for detecting, with respect to each of the change point candidates for the plurality of time series, a global change point that is a change point for the plurality of time series based on a difference between a parameter of a first partial time series starting from a time point of a corresponding change point candidate and a parameter of a second partial time series before the corresponding change point candidate, and outputting the global change point.

Abstract: A computer-implemented method for monitoring computing system status by implementing a deep unsupervised binary coding network includes receiving multivariate time series data from one or more sensors associated with a system, implementing a long short-term memory (LSTM) encoder-decoder framework to capture temporal information of different time steps within the multivariate time series data and perform binary coding, the LSTM encoder-decoder framework including a temporal encoding mechanism, a clustering loss and an adversarial loss, computing a minimal distance from the binary code to historical data, and obtaining a status determination of the system based on a similar pattern analysis using the minimal distance.

Abstract: A method for embedding learning and clustering for paired multi-modal data using deep canonical correlation analysis and active learning with pairwise queries is presented. The method includes collecting time-series data from a plurality of sensors, training, in an unsupervised manner, a cross-modal retrieval system by using the time-series data and relevant comment texts, depending on a modality of a query, retrieving the relevant comment texts from a time-series segment of the time-series data, the relevant comment texts used as human-readable explanations of a query segment, retrieving relevant time-series segments given a sentence or a set of keywords such that the relevant time-series segments match the sentence or set of keywords, and retrieving the relevant time-series segments given the time-series segment and the sentence or set of keywords such that a first subset of attributes match the set of keywords and a second subset of attributes resembles the time-series segment.

Abstract: A system for cross-modal data retrieval is provided that includes a neural network having a time series encoder and text encoder which are jointly trained using an unsupervised training method which is based on a loss function. The loss function jointly evaluates a similarity of feature vectors of training sets of two different modalities of time series and free-form text comments and a compatibility of the time series and the free-form text comments with a word-overlap-based spectral clustering method configured to compute pseudo labels for the unsupervised training method. The computer processing system further includes a database for storing the training sets with feature vectors extracted from encodings of the training sets. The encodings are obtained by encoding a training set of the time series using the time series encoder and encoding a training set of the free-form text comments using the text encoder.

Abstract: A system for cross-modal data retrieval is provided. The system includes a database for storing training sets of two different modalities of time series and free-form text comments as pairs of mixed modality data. The computer processing system further includes a neural network having a time series encoder and text encoder which are jointly trained using a canonical correlation analysis that finds transformations of feature vectors from among the pairs of mixed modality data such that correlated mixed modality data is emphasized in the two different modalities and uncorrelated mixed modality data is minimized. The feature vectors are obtained by encoding a training set of the time series using the time series encoder and encoding a training set of the free-form text comments using the text encoder.

Abstract: A system for cross-modal data retrieval is provided which includes a neural network having a time series encoder and text encoder jointly trained based on a triplet loss relating to two different modalities of (i) time series and (ii) free-form text comments. A database stores training sets with feature vectors extracted from encodings of the training sets. The encodings are obtained by encoding the time series using the time series encoder and encoding the text comments using the text encoder. A processor retrieves the feature vectors corresponding to at least one of the modalities from the database for insertion into a feature space together with a feature vector corresponding to a testing input relating to at least one of a testing time series and a testing free-form text comment, determines a set of nearest neighbors from among the feature vectors based on distance criteria, and outputs testing results.

Abstract: Systems and methods for updating a classification model of a neural network are provided. The methods include selecting, as a set of landmarks, a limited number of data from a set of historical data used to train a classification model. Additionally, the methods generate new training data from recently collected data. Further, the methods update the classification model with the new training data and the set of landmarks to obtain an updated classification model having a loss function configured to capture similarities in the new training data and remember similarities in the historical data represented by the set of landmarks within a predefined tolerance.

Abstract: Factors, other than external factors, that affect the state change of a system are correctly specified, even when an external factor having a strong correlation with the state change of the system exists and preprocessing is required in the time series of an analysis object. An analysis device 500 includes a time series acquisition unit 510, a feature-time series conversion unit 520, and an influence degree calculation unit 540. The time series acquisition unit 510 acquires a difference time series between the value of a target time series and a predicted value of the target time series generated on the basis of the value of a first description time series in a plurality of description time series, and a second description time series. The feature-time series conversion unit 520 extracts, for each of the second description time series, a feature quantity from the second description time series and converts the feature quantity into a feature time series.

Abstract: A factor analysis device includes: a grouping unit 501 to divide a plurality of time-series of explanation that are time-series of data of a plurality of explanatory variables corresponding to a time-series of objective that is time-series of data of one objective variable, into one or more groups such that time-series of explanation having a similarity relationship belong to a same group; a representative time-series extraction unit 502 to extract a representative time-series of explanation from each group; and an analysis unit 503 to analyze an extracted time-series of explanation to identify a time-series of explanation considered to be an influence factor for the time-series of objective.

Abstract: A factor, other than an external factor, having an influence on a state change of a system can be correctly identified even when an external factor having a strong correlation with the state change of the system exists. In an analysis system 1, an external factor identification unit 310 identifies a first explanatory time series among a plurality of explanatory time series. A differential time series generation unit 340 generates a difference time series between a value of an objective time series and a prediction value of the objective time series calculated based on a value of the first explanatory time series. An effect degree calculation unit 420 calculates, based on second explanatory time series among the plurality of explanatory time series and the difference time series, an influence degree of each of the second explanatory time series on a value change of the difference time series.

Abstract: Provided is a factor analysis device capable of obtaining more useful knowledge relating to the degree of influence of pieces of data. A factor analysis device according to one embodiment of the present invention is provided with: a classification unit for classifying a type of data into a first group or a second group; and an influence degree calculation unit for calculating, as the degree of influence on target data, the degree of influence of the data of the type classified into the second group on the data of the first group type.

Abstract: A system monitoring apparatus has: a determination unit for determining whether the relatedness indicating a relationship that is true for first time-series data of multiple sets measured during a first time period is true for second time-series data of the multiple sets measured during a second time period; an irregularity degree calculation unit for calculating the degree of irregularity indicating the extent to which the second time-series data is irregular; a first extent calculation unit for calculating a first extent indicating the extent to which the irregularity is at a specified value when the second time-series data are normal or irregular; a second extent calculation unit for calculating a second extent indicating the extent to which the second time-series data of the multiple sets are related; and a state calculation unit for finding whether the second time-series data are normal or irregular, based on the first and second extents.

Abstract: A display control apparatus includes the following. A first display controller controls a display to display a schedule in which predetermined date information or time span information is corresponded with each of a plurality of schedule frames in a schedule template in which the plurality of schedule frames are provided in a predetermined format. A second display controller controls the display to display a predetermined operation button so that the first display controller displays a plurality of schedules in an aligned state. When the first display controller displays the plurality of schedules in the aligned state according to operation of the predetermined operation button, a portion of each schedule is displayed.

Abstract: This factor analysis device is provided with a feature extraction unit (1021) that extracts feature quantities from an explanatory time series, a feature conversion unit (1022) that converts said feature quantities to a feature time series, a feature-time-series influence-degree computation unit (1031) that uses said feature time series and a response time series to compute an influence degree indicating the degree to which the feature time series influences the change over time represented by the response time series, and an explanatory-time-series influence-degree computation unit (1032) that uses said influence degree to compute an influence degree indicating the degree to which the explanatory time series influences the change over time represented by the response time series.