Hewlett Packard Enterprise Development LP

Hewlett-packard Development Company, L.p.

Netapp, Inc.

Gunalan Perumal Vijayan

Satish Kumar Mopur

Sridhar Balachandriah

Krishnaprasad Lingadahalli Shastry

Saikat Mukherjee

Ashutosh Agrawal

Gregory S Battas

Krishna Prasad Lingadahalli Shastry

William D Dallas

Bhaskar Singhal

Vivek Mehrotra

KrishnaPrasad Lingadahalli Shastry

Satyaprakash Rao

Suresh Ladapuram Soundarajan

Suresh Ladapuram Soundararajan

Systems and methods are provided for implementing a Siamese neural network using improved 'sub' neural networks and loss function. For example, the system can detect a granular change in images using a Siamese Neural Network with Convolutional Autoencoders as the twin sub networks (e.g., Siamese AutoEncoder or “SAE”). In some examples, the loss function may be an adaptive loss function to the SAE network rather than a contrastive loss function, which can help enable smooth control of granularity of change detection across the images. In some examples, an image separation distance value may be calculated to determine the value of change between the image pairs. The image separation distance value may be determined using an Euclidean distance associated with a latent space of an encoder portion of the autoencoder of the neural networks.

Enabling feature importance using Siamese autoencoders for effective image change detection

The present invention relates to a system and a method for updating data models. Input data received from a data source and/or prediction data obtained from a data model is reduced based on baseline reference data to obtain a plurality of representative points. The plurality of representative points are clustered to generate a plurality of clusters. An outlier cluster is detected from the plurality of clusters based on a maximum distance of the plurality of clusters from a highest density cluster and/or comparison of quantity and values of the plurality of representative points with predefined rules. Data drift is identified based on changes in densities of the plurality of clusters. The data model is updated using information corresponding to the outlier cluster and the data drift.

Updating data models to manage data drift and outliers

Systems and methods are provided for retraining machine learning (ML) models. Examples may automatically identify skewed, anomalous, and/or drift occurrence data in real-world input data. By automatically identifying such data, examples can reduce subjectivity in ML model retraining as well as reduce time spent determining a need to retrain a ML model. Accordingly, a determination can be made objectively by a computing system or device according to computer-implemented instructions. Additionally, examples may automatically isolate and transfer data relevant to the retraining of a ML model to a training environment for retraining the ML model using real-world input data. Examples also synthesize large samples of data for use in retraining a ML model. The synthesized data may be generated based on the isolated and transferred data and can be used in place of actual real-world input data to reduce a corresponding delay.

Image data synthesis using adversarial autoencoders for continual training of models

The disclosure relates to a framework for dynamic management of analytic functions such as data processors and machine learned ('ML') models for an Internet of Things intelligent edge that addresses management of the lifecycle of the analytic functions from creation to execution, in production. The end user will be seamlessly able to check in an analytic function, version it, deploy it, evaluate model performance and deploy refined versions into the data flows at the edge or core dynamically for existing and new end points. The framework comprises a hypergraph-based model as a foundation, and may use a microservices architecture with the ML infrastructure and models deployed as containerized microservices.

Intelligent lifecycle management of analytic functions for an IOT intelligent edge with a hypergraph-based approach

A system and method for accounting for the impact of concept drift in selecting machine learning training methods to address the identified impact. Pattern recognition is performed on performance metrics of a deployed production model in an Internet-of-Things (IoT) environment to determine the impact that concept drift (data drift) has had on prediction performance. This concurrent analysis is utilized to select one or more approaches for training machine learning models, thereby accounting for the temporal dynamics of concept drift (and its subsequent impact on prediction performance) in a faster and more efficient manner.

Systems and methods for determining machine learning training approaches based on identified impacts of one or more types of concept drift

Intelligent lifecycle management of analytic functions for an IoT intelligent edge with a hypergraph-based approach

The disclosure relates to technology that implements flow control for machine learning on data such as Internet of Things ('IoT') datasets. The system may route outputs of a data splitter function performed on the IoT datasets to a designated target model based on a user specification for routing the outputs. In this manner, the IoT datasets may be dynamically routed to target datasets without reprogramming machine-learning pipelines, which enable rapid training, testing and validation of ML models as well as an ability to concurrently train, validate, and execute ML models.

Intelligent IOT data split management

A request that indicates a component that can be quiesced is received at a first node. It is determined that quiescence of the component might impact an endpoint. A request for identification of at least one path between a second node and the endpoint is sent to the second node. It is determined, based on a response received from the second node, whether the endpoint will be accessible to the second node if the component is quiesced. In response to a determination that the endpoint will be accessible to the second node if the component is quiesced, a positive analysis outcome is indicated. In response to a determination that the endpoint will not be accessible to the second node if the component is quiesced, a negative analysis outcome is indicated.

Clustered storage system path quiescence analysis

A method, system and computer program product for migrating at least one switch in a storage area network is disclosed. The migration is done by analysing the I/O traffic to identify patterns in the I/O traffic of the switch; forecasting future I/O workload of the switch based on one or more identified patterns in the I/O traffic, determining appropriate timing for migration based on the identified patterns and administrator inputs; processing the storage area network configuration data to identify the storage network physical and logical access paths to the or each selected switch to create a first connectivity map; generating a second connectivity map based on the first connectivity map and administrator inputs; and migrating the or each switch migration based on the second connectivity map and the appropriate timing. The migration may comprise routing the I/O traffic from the switch to be migrated to the alternate switches in the storage area network. The migration may further comprising transforming zones on the switch to be migrated and deploying the transformed zones to the new switch.

Growing community of inventors

Bengaluru, India

Gunalan Perumal Vijayan