Project Background
The AI/ML Enhancement Project innovatively integrates cutting-edge Artificial Intelligence (AI) and Machine Learning (ML) technologies into the Geohazards Exploitation Platform (GEP) and Urban Thematic Exploitation Platform (U-TEP) through MLOps - the fusion of ML with DevOps principles (see Project Launch Article - May 2023).
The project’s objective was to enhance the value proposition of these platforms, expanding the range of services by combining EO data exploitation with AI/ML model development, deployment, and execution. These new capabilities were tested from the perspective of two types of users. This testing was conducted across 10 User Scenarios, designed to reflect the key activities and goals of each user type in utilising the service.
User Personas
- “Alice”, a ML Practitioner who is an expert in building and training ML models, selecting appropriate algorithms, analysing data, and using ML techniques to solve real-world problems.
- “Eric”, a Consumer who is a stakeholder or user (e.g. business owner, a customer, a researcher, etc) who benefits from, or relies upon, the insights or predictions generated by the ML models to inform his decision-making process.
User Scenarios
- User Scenario 1 - Alice does Exploratory Data Analysis (EDA)
- User Scenario 2 - Alice labels Earth Observation data
- User Scenario 3 - Alice describes the labelled Earth Observation data
- User Scenario 4 - Alice discovers labelled Earth Observation data
- User Scenario 5 - Alice develops a new Machine Learning model
- User Scenario 6 - Alice starts a training job on a remote machine
- User Scenario 7 - Alice describes her trained machine learning model
- User Scenario 8 - Alice reuses an existing pre-trained model
- User Scenario 9 - Alice creates a training dataset
- User Scenario 10 - Eric discovers a model and consumes it
Project Status and Outreach Articles
The development phase of the Project is now completed, with the successful implementation of all User Scenarios. The functionalities of the EO data exploitation with AI/ML model development, deployment, and execution were tested through the execution of Jupyter Notebooks and Application Packages created in each User Scenario, according to its specific requirements.
The user manual for the new functionalities is available on Github at https://ai-extensions.github.io/docs/. Here, users can access the relevant resources for AI/ML model development, deployment, and execution.
Furthermore, each User Scenario was documented in a dedicated outreach article, illustrating the coding resources and procedures made available to users through its implementation. Below is a list of these articles with links to their publications:
- Article on User Scenario 1 “Exploratory Data Analysis”
- Article on User Scenario 2 “Labelling EO Data”
- Article on User Scenario 3 “Describing labelled EO data”
- Article on User Scenario 4 “Discovering labelled EO data with STAC”
- Article on User Scenario 5 “Developing a new ML model and tracking with MLflow”
- Article on User Scenario 6 “Training and Inference on a remote machine”
- Article on User Scenario 7 “Describing a trained ML model”
- Article on User Scenario 8 “Reusing an existing pre-trained model”
- Article on User Scenario 9 “Creating a training dataset”
- Article on User Scenario 10 “Discovering, Deploying and Consuming an ML model”
What’s next?
The coding resources developed through this project are designed to be executed from within the platform’s service integration environment, after a proper user onboarding process that configures the platform to work correctly. To request account creation and configuration, please send an email to support@terradue.com with subject “Request Access to Notebook X” and body “Please provide access to Notebook X”.
Want to get involved? We invite anyone interested in learning more about the AI/ML model development, deployment, and execution service and the related coding resources to reach out to us!
As we enter the project’s engagement phase, we’re fostering discussions through our community forum on Terradue’s Discuss website, where you can ask questions, share feedback, and connect with both the project team and other users. Join us in shaping the future of this platform and exploring how it can benefit your work in EO and AI.