Radiant MLHub
- Washington DC, United States of America
- December 2019
- Radiant Earth Foundation
Radiant MLHub is a cloud-based open library dedicated to open geospatial training datasets and models generated by Radiant Earth Foundation, partners, and the community. Radiant MLHub allows anyone to access, store, register, and share open training datasets and models for high-quality Earth observations, and it’s designed to encourage widespread collaboration and the development of trustworthy applications.
- Open source
- 64/35/- (% m/f/d)
- Both public and private (Foundation Awards, Government Grants and Commercial Contracts) funding
Project stage (in ):
Implemented/Running Project
Implemented by:
Inhouse
Industrial Sectors:
Agriculture, forestry and fishing, Professional, scientific and technical activities, Other service activities, Earth Observation, Climate Change
Usage of AI:
Computer Vision, Data Management and Analysis, Information Retrieval, Generative Models
Generation of AI:
Traditional Machine Learning (Linear Regression, CART, SVM, etc.), Deep Learning (CNN, Transformers, etc.)
Model training:
Supervised Learning, Semi-supervised Learning, Transfer Learning
Contact
Responsible Person
Hamed Alemohammad
Motivation and values
What in your view defines the public interest and how does your project meet this purpose?
Earth observation (EO) and Machine Learning (ML) can play a major role across the Global Development Community (GDC) in working to achieve UN Sustainable Development Goals by providing consistent data and monitoring tools. Over the last decade, the volume of EO data has been dramatically increasing, enabling products and applications across markets from agriculture to medicine to transportation. EO and ML present a game-changing opportunity to more accurately and quickly identify and address unique, complex, and emerging challenges at local, regional, and global scales. However, multiple issues should be addressed for the GDC to benefit from these technologies and contribute to their advancements for global missions. Lack of geodiversity in training datasets and difficulty in discovering and accessing datasets and models are among the high-priority challenges to enable development of unbiased models especially for problems in developing regions. Furthermore, generating high-quality training data requires extensive effort and coordination from practitioners.
How did the idea of your project come about?
Internal in our organization
What is the goal of your project in relation to the public interest?
The goal of Radiant MLHub is to generate and aggregate open geospatial training data and models for AI practitioners and researchers working on international development challenges. This will result in increased availability of high-quality open training data and models, enhanced geodiversity of data, improvements in models’ accuracy, and expansion in the diversity of EO applications. We envision that this data ecosystem will facilitate innovation across the global development sector, and enable more efficient product development cycles. Practitioners and policy makers working on global development challenges will have high-quality and transparent data and products that they can trust to support evidence-based decisions. Ultimately, poor and vulnerable people in the Global South will benefit from improved living conditions and economic opportunity because of actionable insight derived from EO that is empowered by community access to high-quality open training data and models.
Did you follow one or more guidelines for ethical AI, and if yes, which one?
FAIR Principles
Guiding Values
What are your top 5 guiding-values for the project? Top 1
Open
What measures do you use to implement this value? Top 1
Radiant MLHub has custom built machine learning tools, libraries and new metadata standard development like SpatioTemporal Asset Catalog (STAC).
What are your top 5 guiding-values for the project? Top 2
Innovative
What measures do you use to implement this value? Top 2
Radiant MLHub has custom built machine learning tools, libraries and new metadata standard development like STAC.
What are your top 5 guiding-values for the project? Top 3
Trust
What measures do you use to implement this value? Top 3
Radiant MLHub is a trusted resource for the community to access benchmarked data, models and standards. The technology that Radiant MLHub is built on, as well as the data and models are transparent.
What are your top 5 guiding-values for the project? Top 4
Collaborative
What measures do you use to implement this value? Top 4
Radiant MLHub is built on a collaborative ecosystem. We sponsor Technology Fellowships to support innovative technology standards such as the STAC specification. We also convene workshops and technology sprints to collaborate on training data standards and best practices.
What are your top 5 guiding-values for the project? Top 5
Diversity
What measures do you use to implement this value? Top 5
Radiant MLHub brings different voices together to advance common goals. Our diversity is reflected within the team, data and applications.
Design & Safeguards
Which stakeholders were involved in the process of development and implementation of the project?
Developer, External domain expert, Civil society, Public administration, Academic researcher, Industry partners
How did you engage relevant stakeholders?
Workshops, Survey, Interviews, Co-creation process
Did you apply specific methods of participatory design, and if yes, which ones?
Professional Collaboration, User Research & Testing
Have the project results been validated by third parties?
Yes
Have the design and the results of the project been made transparent to the public?
Yes
In what way have the design and the results of the project been made transparent to the public?
By a public organization, By academic researchers, By the open-source community
Involving the people who will be affected by the project is a necessary part of the project design
5
Which direct channels for feedback exist?
Public Slack workspace, GitHub & Email
How did you verify the accuracy and robustness of your system?
User testing
Which transparency measures do you use to document your use of data and possible biases or limitations of the data sets?
Data sheets
What technical and organizational safeguards have you implemented to protect personal data and mitigate possible harms? Choose all that apply
Data minimization (incl. not gathering personal data), Authentication and access management, Maintaining an up to date privacy policy
Do you take measures in regards to ecological sustainability of your project? Please elaborate.
Not at this time.
Is your project directly or indirectly contributing to solutions for the climate crisis or other UN sustainability goals?
Directly
Please elaborate
Radiant MLHub supplies AI practitioners and researchers working on sustainable development goals such as climate change, food security, natural resource management with high-quality geospatial training data and machine learning models. This allows practitioners to create solutions for the problem(s) they are working on. The datasets and models are geographical diverse to help practitioners build more accurate applications and deploy the products to the market to improve evidence-based decision making.
Does your project rely partly or fully on the use of open data and/or publish results in an open data set? Please elaborate.
Yes, Radiant MLHub is an open-access geospatial training data repository that anyone can use to discover and download ML-ready training datasets. The concept of open is reflected in our mission statement which is “to empower organizations and individuals with open EO and ML tools and datasets to better address international development challenges.” We also believe that “open is not enough” and that the global development community needs better access to high-quality standardized data that is reachable and usable for a variety of applications.