Emc IP Holding Company LLC

Emc Corporation

Diego Salomone Bruno

Percy E Rivera Salas

Rômulo Teixeira De Abreu Pinho

Tiago Salviano Calmon

Jonas Furtado Dias

Owen Martin

Vinícius Michel Gottin

Adriana Bechara Prado

André De Almeida Maximo

Rodrigo Dias Arruda Senra

Victor Bursztyn

Monica Barros

Caio Souza

Techniques are provided for adaptive look-ahead configuration for data prefetching. One method comprises, in response to a request for a data item in a storage system: obtaining a size of a look-ahead window for the request based on one of multiple available caching policies; and moving the requested data item and additional data items within the look-ahead window to the cache memory when the requested data item and/or the additional data items within the look-ahead window are not in the cache memory. The multiple available caching policies comprise a caching policy based on characteristics of an input/output workload of the storage system, or a portion thereof; and/or a caching policy based on an input/output workload of at least a portion of the storage system within a prior predefined time window. The look-ahead window size may be varied over time.

Adaptive look-ahead configuration for prefetching data in input/output operations

Relevance decay techniques are provided for time-based evaluation of machine learning applications and other classifiers. An exemplary method comprises obtaining time series measurement data; generating an input dataset comprising a plurality of records, wherein each record comprises features extracted from the time series measurement data, a target class corresponding to an event to be identified, and a time lag indicating a difference in time between a given extraction and the event to be identified; evaluating a plurality of classifiers during an evaluation phase using a portion of the input dataset and one or more predefined evaluation metrics weighted using a time-based relevance decay function based on the time lag; and selecting one or more of the classifiers to perform classification of the time series measurement data based on the predefined weighted evaluation metrics during a classification phase. The time lags indicate, for example, a time difference between classification moments of the plurality of classifiers and a respective instance of the event to be identified.

Relevance decay for time-based evaluation of machine learning applications

Traffic condition forecasting techniques are provided that use matrix compression and deep neural networks. An illustrative method comprises obtaining a compressed origination-destination matrix indicating a cost to travel between pairs of a plurality of nodes, wherein the compressed origination-destination matrix is compressed using a locality-aware compression technique that maintains only non-empty data; obtaining a trained deep neural network trained using the compressed origination-destination matrix and past observations of traffic conditions at various times; and applying traffic conditions between two nodes in the compressed origination-destination matrix at a time, t, to the trained deep neural network to obtain predicted traffic conditions between the two nodes at a time, t+&#x394;. A tensor can be generated indicating an evolution of traffic conditions over a time span using a stacked Origination-Destination matrix comprising a plurality of past observations representing the time span.

Traffic condition forecasting using matrix compression and deep neural networks

Methods and apparatus are provided for pattern-aware transformation of time series data to multi-dimensional data for Deep Learning analysis. An exemplary method comprises: obtaining time series data and an indication of seasonal components in the time series data; obtaining the time series data separated into data chunks of a predefined length based on at least one seasonal component; aligning the data chunks based on the at least one seasonal component; generating an image and/or a multi-dimensional vector using the aligned data chunks; and applying the image and/or the multi-dimensional vector to a Deep Learning module to obtain a prediction, a classification and/or a profiling of parameters associated with the time series data. The classification of the parameters comprises, for example, an anomaly detection. The profiling of the parameters comprises, for example, a clustering of the parameters and/or a behavior identification.

Pattern-aware transformation of time series data to multi-dimensional data for deep learning analysis

Event data collected for a given event is obtained, wherein the event data comprises a plurality of time series data sets. The plurality of time series data sets are divided into a set of time windows epochs). Data in the plurality of time series data sets occurring within each time window of the set of time windows is aligned. A metric is computed for each aligned time window, wherein the metric for each aligned time window represents a measure of at least one of completeness and support attributable to data in the aligned time window. Data is pruned from the set of event data for one or more of the set of time windows based on the computed metrics. The pruned event data is provided to a data analytics process which is configured to further process the pruned event data.

Decision processing applied to data analytics workflow

Methods and apparatus are provided for real-time ad hoc querying of data records, such as call detail records. An exemplary method comprises obtaining data records, wherein each data records comprises a plurality of fields and wherein one or more fields are used to aggregate a history of the data records; storing the data records in a consolidation queue (and optionally, a waiting queue), based on predefined indexing criteria; storing the data records from the consolidation queues in a historical database based on one or more predefined criteria related to a queue length and/or a time interval; creating a pre-processed index that maps indexed information for a given entity to a portion of the historical database; and, in response to a user query regarding attributes of the given entity: retrieving the data records related to the given entity from the historical database based on the pre-processed index for the given entity; storing the retrieved data records of the given entity into an in-memory database; and processing the retrieved data records of the given entity to answer the one or more user queries.

Real-time ad hoc querying of data records

Methods and apparatus are provided for domain-tailored detection of outliers, patterns, and/or events in data streams. An exemplary method comprises obtaining a domain-dependent definition of (i) data outliers based on predefined outlier criteria; (ii) data patterns based on predefined pattern criteria; and/or (iii) data events based on predefined event criteria; obtaining time series measurement data from a plurality of sensors; determining, substantially simultaneously with the obtaining, whether individual samples satisfy the domain-dependent definitions of the data outliers, data patterns and/or data events; and storing the individual samples with an indication of whether the individual samples satisfy the domain-dependent definitions of the data outliers, data patterns and/or data events. The domain-dependent definitions are optionally specified using a declarative command language. Query are optionally processed comprising one or more declarative statements that reference and/or manipulate the data outliers, data patterns and/or data events.

Database system and methods for domain-tailored detection of outliers, patterns, and events in data streams

Event data collected for a given event is obtained, wherein the event data comprises a plurality of time series data sets. The plurality of time series data sets are divided into a set of time windows (e.g., epochs). Data in the plurality of time series data sets occurring within each time window of the set of time windows is aligned. A metric is computed for each aligned time window, wherein the metric for each aligned time window represents a measure of at least one of completeness and support attributable to data in the aligned time window. Data is pruned from the set of event data for one or more of the set of time windows based on the computed metrics. The pruned event data is provided to a data analytics process which is configured to further process the pruned event data.

Growing community of inventors

Niterói, Brazil

Diego Salomone Bruno