KI:STE
- Jülich, Germany
- November 2020
- Universities in Aachen (RWTH), Bonn and Cologne together with Forschungszentrum Jülich, and the companies 52°N and Ambroys
The project aims to exploit recent developments in artificial intelligence for environmental data analysis. The goal is the implementation of current AI approaches for spatiotemporal variable pattern recognition and pattern analysis in environmental data (for clouds, natural hazards, water, air quality and vegetation). A technical platform will be created to make powerful AI applications on environmental data available in a portable way. An online AI learning platform will be established. This e-learning offers location-independent education of young scientists. It will use the concepts and methods developed in the five research fields as teaching material.
- Open source
- 57/43/- (% m/f/d)
- Public
Project stage (in ):
Beta/Testing
Implemented by:
Inhouse, External
Industrial Sectors:
Agriculture, forestry and fishing, Water supply; sewerage, waste management and remediation activities, Natural Hazards
Usage of AI:
Computer Vision, Generative Models
Generation of AI:
Deep Learning (CNN, Transformers, etc.)
Model training:
Supervised Learning, Semi-supervised Learning
Contact
Responsible Person
Jülich Supercomputing Centre
Motivation and values
What in your view defines the public interest and how does your project meet this purpose?
We taylor AI to support solving environmental problems. With our AI we tackle problems such as natural hazards (land slides), environmental protection (wilderness) and cloud classification for improved solar power plants. The motivation behind the project is to adapt AI to environmental data, which has many applications beneficial for humanity.
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?
Besides tackling environmental challenges with AI, we also communicate our insights to other environmental researchers and master students. The e-learning platform is online and free so that the public could also check out our educational material.
Did you follow one or more guidelines for ethical AI, and if yes, which one?
European Union - Ethics Guidelines for Trustworthy AI
Guiding Values
What are your top 5 guiding-values for the project? Top 1
Open Science
What measures do you use to implement this value? Top 1
Publish in open journals, publish our code on Git for download and re-use
What are your top 5 guiding-values for the project? Top 2
Good Scientific Practice
What measures do you use to implement this value? Top 2
Peer-Reviewing, Pair Programming, proper education of our doctoral researchers
What are your top 5 guiding-values for the project? Top 3
Trustworthy AI
What measures do you use to implement this value? Top 3
Overcome the black-box nature by carefully desgining and explaining our AI.
What are your top 5 guiding-values for the project? Top 4
FAIR Data
What measures do you use to implement this value? Top 4
Clear documentation and bridges to accessible, open data via our AI-platform
What are your top 5 guiding-values for the project? Top 5
Tackle Real World Problems
What measures do you use to implement this value? Top 5
Instead of developing AI for the purpose of having AI even better at recognizing cats and dogs, we tackle the highly complex spatio-temporal environmental data to solve real-world problems
Design & Safeguards
Which stakeholders were involved in the process of development and implementation of the project?
Developer, Academic researcher, Industry partners
How did you engage relevant stakeholders?
Workshops, Co-creation process
Did you apply specific methods of participatory design, and if yes, which ones?
I do not know.
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 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?
Direct and close communication within the project partners. Moreover academic peer review.
How did you verify the accuracy and robustness of your system?
Besides "traditional" machine learning evaluation metrics, we implement domain-specific evaluation metrics to make sure the accuracy is conisistent with our prior knowledge. Thus, accuracy and robustness is tested twice.
Which transparency measures do you use to document your use of data and possible biases or limitations of the data sets?
Documentation
What technical and organizational safeguards have you implemented to protect personal data and mitigate possible harms? Choose all that apply
Do you take measures in regards to ecological sustainability of your project? Please elaborate.
None.
Is your project directly or indirectly contributing to solutions for the climate crisis or other UN sustainability goals?
Indirectly
Please elaborate
By developing AI able to analyze environmental data and tackle environmental problems.
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, open data is used, the code is published openly. In case we heavily preprocessed the data, the preprocessing routines are made availbale or the dataset is published as benchmark dataset.