ML Model Publisher Automates Deployment via Self-Service Interface
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing process of publishing machine learning models is time-consuming, labor-intensive, and requires manual coding, especially when deploying models to non-AZURE targets.
Innovation Solution
A single-click ML model publisher that runs within a ML modeling application on an AI platform, allowing users to publish ML models without packaging or additional coding by generating a publication request form and accessing data structures in memory to populate the form with necessary attributes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manual coding is used to package and deploy ML models to non-AZURE targets, then the model can be deployed to various target systems, but the process becomes time-consuming and labor-intensive
Solution Approach 1:
The system enables self-service deployment by automatically generating deployment configurations and container images without requiring manual coding. The ML model publisher automatically queries the data structure, populates the publication request form, and executes the deployment process, allowing users to deploy models with a single click without manual intervention in the packaging or configuration steps.
Solution Approach 2:
The system performs preliminary actions by pre-configuring the deployment pipeline and pre-populating the publication request form with data from the data structure. The container image template and deployment configuration are prepared in advance, so when deployment is initiated, the system only needs to execute the pre-prepared configuration rather than requiring manual coding at deployment time.
2Adaptability or versatility
If manual coding is required for model packaging, then the model can be customized for different targets, but the process becomes tedious and requires in-depth programming knowledge
Solution Approach 1:
The system eliminates the need for manual coding by implementing self-service automation. The ML model publisher automatically queries the data structure to obtain model attributes, generates the appropriate container image configuration, and executes the deployment to the target system. Users simply need to initiate the publication process through the user interface, and the system handles all technical details automatically.
Solution Approach 2:
The system introduces an intermediary layer in the form of the ML model publisher and the publication request form. This intermediary automatically translates the user's deployment intent into the necessary technical configurations and executes the deployment process. The intermediary handles the complex tasks of container image generation and deployment configuration, shielding users from the underlying technical complexity.
3Adaptability or versatility
If traditional packaging processes are used, then the model can be deployed to various environments, but the process requires in-depth programming knowledge and knowledge of the target system
Solution Approach 1:
The system implements self-service automation that eliminates the need for users to possess in-depth programming knowledge. The ML model publisher automatically handles all technical aspects including container image generation, dependency management, and deployment configuration. Users simply interact with the user interface to initiate deployment, and the system automatically queries the data structure, generates the necessary configurations, and executes the deployment to the target environment.
Solution Approach 2:
The system replaces the mechanical process of manual coding and configuration with an automated software-based system. Instead of requiring users to write code and manually configure deployment parameters, the system uses automated software components including the ML model publisher, publication request form, and container image template to handle all technical tasks. This substitution of manual mechanical operations with automated software processes eliminates the need for in-depth programming knowledge while maintaining deployment flexibility.
Data Source
AI summary
A machine learning (ML) model publisher can, responsive to an indication that a ML model is ready for publication, generate a publication request form or page on a user device. The ML model publisher can be invoked from within a ML modeling application. Responsive to an instruction received through the publication request form or page, the ML model publisher can access a data structure in memory used in training the ML model and populate the publication request form or page with attributes required by the ML model to run. Responsive to activation of a single-click publication actuator, the ML model publisher can publish the ML model directly from the ML modeling application to a target computing system by providing, to the target computing system, a path to a repository location where the ML model is stored and information on the attributes required by the ML model to run.