The network plans to be fully operational by early 2025. The goals of 2024 are setting up its
organization and looking for suitable projects and partners as well as running pilot projects to
generate early impact and learnings for the setup.
1. Improved Weather Prediction for Sustainable
Agriculture
In order to improve agricultural production in Africa, it is important to improve the accuracy of
weather forecasts available to small scale farmers. Agriculture in Africa is mainly rain fed and
accurate rainfall prediction is likely to improve yields by allowing farmers to appropriately time
activities such as planting and also help them select appropriate crops to grow.
Goal: In this project we will leverage weather data collected by large networks of sensors to
improve weather prediction by linking these data with new artificial intelligence based weather
prediction models running on powerful computers. The key goal is to provide accurate medium to long
range localised precipitation forecasts.
Partners: This work will bring together
Data
Science Africa (DSA), the
Centre for Data Science
and
Artificial Intelligence (DSAIL) at Dedan Kimathi University of Technology in Kenya,
Amini, a Kenyan climate startup and
Communal Shamba Coffee.
2. Early diagnosis of plant diseases through
spectroscopy
About 40% of the global crop production is lost to pests. Sub-Saharan Africa is most vulnerable to
the increasing risks of pests and diseases spreading in agriculture. The current methods of disease
identification and diagnosis involve experts traveling to disparate parts of the country and
visually scoring the plants by looking at the disease symptoms manifested on the leaves.
Goal: The research project investigates a 3-D printed smartphone add-on spectrometer that
determines the state of disease in plants before it is visibly symptomatic. Portable devices for the
early detection of crop diseases are needed to support the farmers and experts working in the field.
The output of this tool is integrated into a smartphone in the form of an app, making it accessible
for use in the field in real applications.
Partners: This project will be executed by
Data
Science Africa (DSA), and
ETH Zurich.
3. Ethical AI for Humanitarian Action - Developing
Tailored LLMs
The International Committee of the Red Cross (ICRC) seeks to leverage Large Language Models (LLMs)
to enhance its humanitarian work. Challenges like the bias of existing models, the
underrepresentation of humanitarian contexts in commercial AI training sets, and the sensitivity of
data related to conflicts limit the adoption of off-the-shelf AI models.
Goal: To address these constraints, the ICRC will investigate the responsible development of
LLMs specifically tailored to its mandate on international humanitarian law and in line with its
protection work in favor of people affected by armed conflicts. Ethical and accountability
considerations will be at the center of this project, with the development of tools to assess if
developed models are safe to be used for humanitarian applications.
Partners: This project will be executed by
ICRC,
ETH Zurich,
and
EPFL.