AI REDGIO 5.0 AI Pipeline Designer
Facilitating the design, execution and deployment of AI pipelines for AI and ML applications
Asset Description
The AI Pipeline Designer platform allows users design, experiment with, train, execute and deploy AI models at the edge or on cloud resources. The AI Pipeline Designer platform is addressed towards different types of users (i.e. data scientists, technical users, business users) in terms of AI and ML execution, from executing simple data manipulation functions (e.g. filters and aggregations) and applying Artificial Intelligence models for the manufacturing domain analytics, to creating visualisations and reports to highlight insights extracted from datasets and from analytics processes, as well as exporting the outputs of these analyses through interfaces that can be consumed by other systems.
Features
| Feature | Description |
|---|---|
| Definition of re-usable and customisable data treatment blocks | Data manipulation functions wrapped in configurable pipeline blocks, allowing the user to select the block that offers the desired functionality, parameterise it and combine it with other blocks to generate a pipeline. |
| Definition of re-usable and customisable analytics models blocks based on mainstream algorithms | Blocks offering generic algorithms that can be trained and configured in a more flexible way to allow users implement their own models. |
| Re-usable and customisable data input blocks | Input blocks indicating data assets that an organisation has at its disposal. |
| Re-usable and customisable data output blocks | Output blocks allowing users download the results of their pipeline as a file, retrieve the results via Open APIs, or visualize themthrough custom diagrams that are created and saved with the help of the Visualization Engine. |
| Configuration and validation of AI pipelines | Providing the execution details of a defined pipeline (e.g. schedule-based execution or/and event-triggered execution). |
| Deployment of AI models on Edge devices | Deploy and execute a model at a selected edge device. |
| Update of data analysis pipelines | Revise parts of a configured and finalised AI pipleine without compromising the consistency of the results. |
| Insights into the execution of AI pipelines | Collection and display of metrics about the outcome and different performance aspects of the data analysis pipeline. |
| Registration of externally trained compatible AI models | Users that have trained ML/DL models outside the AI REDGIO 5.0 Platform yet in compatible libraries, can register them in the platform. |
Resources
- Created by SUITE5 Data Intelligence Solutions
- For more information contact here