/7.1/aaw/Recent content in Active Analytics Workbench (AAW) on Kinetica DocsHugo -- gohugo.ioen/7.1/aaw/overview/Mon, 01 Jan 0001 00:00:00 +0000/7.1/aaw/overview/Kinetica provides the Active Analytics Workbench (AAW) with the goal of simplifying and accelerating data science and machine learning in a scalable fashion. With AAW, users can ingest data, train models, make inferences (answers/output from models), and even audit models with a few endpoints (or clicks via the UI). The AAW package can be automatically installed via KAgent and coexists with the database--meaning easy access to one's data and GPUs./7.1/aaw/aaw_ui/Mon, 01 Jan 0001 00:00:00 +0000/7.1/aaw/aaw_ui/ The Active Analytics Workbench (AAW) UI allows for simplified and controlled usage of the Kinetica machine learning capabilities. AAW is usually installed on port 8070 of the Kinetica instance, so it can be accessed via http://<kinetica.aaw.node>:8070/. The AAW UI home page is the Dashboard. Dashboard Data IngestsDatasetsFeature Sets Models + Analytics DetailsModel CreationModel Deployment RAPIDS Notebook Workspace PrerequisitesOverview Deployments Details Audits Security DetailsNew Credential/7.1/aaw/audits/Mon, 01 Jan 0001 00:00:00 +0000/7.1/aaw/audits/Audits provide a means for reviewing inferences usage and accuracy. To initiate an audit search: Select a Model that has been deployed. Select a Date/Time Range. Optionally, add Search Filters: Click + Add Filter as many times as there are filters necessary to refine the search. For each filter, select a column and an operator & provide a value, e.g., fare_amount > 0. Click Search. Should the audit find anything, the Results table will become available./7.1/aaw/dashboard/Mon, 01 Jan 0001 00:00:00 +0000/7.1/aaw/dashboard/The Dashboard is the home page for the AAW user interface. It provides a detailed summary of all entities stored within AAW as well as their status. Click Refresh All to refresh the Overview graphs and the entity tables below. The Overview displays a graph for each entity type: Ingests, Datasets, Models, and Deployments. Each graph is divided into portions that represent n entities that occupy one of six statuses./7.1/aaw/data/Mon, 01 Jan 0001 00:00:00 +0000/7.1/aaw/data/Ingests Details Batch Streaming Import BYOC Datasets Details New Dataset Feature Sets Details New Feature Set Ingests An ingest is a means for getting data into Kinetica to be used in models. Point to + Add New Ingest then click one of the options to begin creating an ingest. If an ingest has an output table, click the output table name to view a preview of the table Type into Filter to filter down the ingests Click to display any archived ingests Click to refresh the table Click to export the table's values as JSON or CSV Click an existing ingest to display additional actions: Click View Selection to open the Ingest Details page Click Start to start the ingest job Click Export Entity to export the ingest as a JSON object Click Terminate to stop the ingest Click Clone to clone the ingest Click Archive to archive the ingest; it will be stopped (if not already) and hidden from the list of ingests Click Description / Config to review summary information for the ingest Details The Ingest Details page provides a detailed look at a given ingest, including configuration information, state, and a data preview (if the ingest has a destination table configured)./7.1/aaw/deployments/Mon, 01 Jan 0001 00:00:00 +0000/7.1/aaw/deployments/A deployment is essentially an active model; if a model is deployed, inferences can be tested against it. After a model has been deployed, an entry for that deployment will be available on the Deployments page. AAW currently supports three types of deployments: On Demand, Batch, and Continuous. On Demand deployments can be tested manually and as necessary with user input. Batch deployments inference on a batch of data all at once./7.1/aaw/blackbox_sdk/Mon, 01 Jan 0001 00:00:00 +0000/7.1/aaw/blackbox_sdk/The Kinetica Blackbox SDK assists users in creating blackbox models to wrap existing code/functionality and make it deployable within the Kinetica system. The Active Analytics Workbench (AAW) currently can only import blackbox models that have been containerized and implement the BlackBox SDK. Users provide the Python module scripts, modify some SDK files, and the SDK will build a Docker Container from the files and publish it to a given Docker Registry (private or public)./7.1/aaw/models_analytics/Mon, 01 Jan 0001 00:00:00 +0000/7.1/aaw/models_analytics/A model is a mathematical or programmatical representation of a real-world process. Point to + Add Model then click one of the options to begin the model setup/import process. New Blackbox -- create a new Blackbox model Type into Filter to filter down the models Click to display any archived models. Click to refresh the table Click to export the table's values as JSON or CSV Click an existing model to display additional actions./7.1/aaw/rapids/Mon, 01 Jan 0001 00:00:00 +0000/7.1/aaw/rapids/The RAPIDS Notebook Workspace is a custom JupyterLab environment integrated within AAW. From the workspace, users can import, create, and develop Jupyter notebooks directly on top of the data necessary for machine learning workflows; monitor GPU processes and hardware; interact with and track in-depth Python workflow analytics; and much more. Prerequisites At least 1 GPU must be available to start a RAPIDS Notebook Workspace session (Kubernetes uses 1 GPU)/7.1/aaw/security/Mon, 01 Jan 0001 00:00:00 +0000/7.1/aaw/security/All credentials stored in AAW are accessible through the Security page. Credentials provide a means for AAW to access various third-party/external resources for ingestion, model, and deployment purposes, e.g., AWS S3, Docker, the Kinetica database, etc. Click Add New Credential to begin the credential setup process Type into Filter to filter down the credentials Click to display any archived credentials. Click to refresh the table Click to export the table's values as JSON or CSV Click an existing credential to display additional actions: Click View Selection to open the Credential Details page Click Export Entity to export the credential as a JSON object Click Archive to archive the credential; it will be hidden from the list of credentials Details The Credential Details page provides a detailed look at a given credential, including login information, pertinent URLs/URIs, and creation information.