ΑΙhub.org
 

AfriClimate AI participation at the Deep Learning Indaba 2024: from a spark to a community, leading AI for climate action

by and
27 September 2024



share this:

The Deep Learning Indaba participants, in Dakar, Senegal 1-7 September 2024.

The Deep Learning Indaba 2024 was not just another event for us—it was a reunion. For AfriClimate AI, the Indaba represents our birthplace. It was at the Indaba 2023 in Accra, Ghana, that a pivotal conversation ignited a movement, sparking the creation of AfriClimate AI. As Rendani Mbuvha, one of our founding members, recalls:

“Last year, I was invited to give a talk about my work on Uncertainty, AI, and Climate Science at the Deep Learning Indaba in Accra, Ghana. As is usual with invited talks, one tends to focus on the successful parts of the work. But for some reason, that morning, I decided to add a slide about the challenges of working in AI and sustainability in Africa, primarily driven by the pervasive data scarcity issues. It turned out that almost everyone in the room identified with these issues. This was the birthplace of AfriClimate AI, a grassroots research community dedicated to tackling these issues head-on through capacity building, open datasets, representative benchmarks, and state-of-the-art weather forecasting models for Africa.”
– Rendani Mbuvha

From that moment in Accra, AfriClimate AI was born out of a shared vision to address Africa’s unique climate challenges using AI. This year, we returned to the Indaba as a well-formed community, ready to showcase our growth and progress.

Eight of the nine core team and founding members of AfriClimate AI.

Day 1 – Tuesday: A homecoming and community session

The first day of Deep Learning Indaba 2024 felt like a homecoming. Walking into the Community Session on Tuesday, we were reunited with sister communities like Masakhane, reminding us of the collaborative spirit that lies at the heart of African AI innovation. We set up our stand, proud to present projects like the “AfriNet” : Weather stations Network project, a bold initiative to build a network of accessible weather stations across the continent and collect and curate the much needed climate datasets. The excitement was contagious, as attendees flocked to learn how our data-driven solutions could be scaled to their own regions.

AfriClimate AI community showcase stand at the Deep Learning Indaba 2024.

This session also marked the beginning of discussions on the data scarcity issue, one of the core challenges that AfriClimate AI aims to tackle. This is where our story began, and now, we return not just with ideas but with practical solutions and a thriving community of researchers, developers, and climate advocates.

Community Showcase Session at the Deep Learning Indaba.

Wednesday: Poster presentation – a key milestone

Before the full-day workshop on Friday, AfriClimate AI reached another significant milestone on Wednesday with our poster presentation. Both John Bagiliko and Amal Nammouchi, core team members, presented a poster that outlined our vision for AI-driven climate action in Africa. The poster focused on key challenges such as data scarcity and the importance of creating open-access datasets to drive climate resilience across the continent.

Our presentation generated a lot of interest, and the paper was not only well-received but also recommended for journal publication. This recognition highlights AfriClimate AI’s growing influence and credibility in the field of AI for climate science, and it was a proud moment for our entire community.

“In June, I had the opportunity to present our position paper in Malmö, Sweden at the HHAI conference, and today, I stand here in Dakar, Senegal, to represent AfriClimate AI once again. It’s a reminder of how far we’ve come as a community, working together to amplify our impact. It’s not just about the technology we create, but the lives we can change by equipping Africa with the tools and knowledge to build climate resilience.”
– Amal Nammouchi

Deep learning Indaba Poster Session, AfriClimate AI position paper.

AfriClimate AI full-day workshop: a day of learning and inspiration

Friday marked the centrepiece of our presence at the Indaba—a full-day workshop that highlighted the power of AI in climate action through a series of keynotes, tutorials, panel discussions, and lightning talks. The day began with an inspiring keynote by Dr Mouhamadou Bamba Sylla, who spoke on the State of Climate Science in Africa and the Opportunities and Risks Presented by Artificial Intelligence (AI).

Dr Sylla, currently serving as the AIMS Network Research Chair in Climate Change Science at the African Institute of Mathematical Sciences in Rwanda, brought his vast experience in climate modelling and change to the stage. Having worked at WASCAL, contributed as a lead author for the Intergovernmental Panel on Climate Change (IPCC), Dr Sylla’s insights underscored the vital role AI can play in extreme weather prediction and climate resilience. His keynote provided an important context for the challenges and opportunities AI presents in the African climate landscape especially when it comes to data accessibility in different African countries. He also highlighted the opportunities that satellite data plays to overcome some of these data issues.

Dr Mouhamadou Bamba Sylla keynote at AfriClimate AI DLI workshop.

Our second keynote speaker, Dr Jason Hickey, delivered a powerful talk on machine learning for weather and climate modelling, emphasising the democratisation of weather forecasting:

“The field is going through a transformational time where ML models have become competitive, making forecasts more available and affordable. I see it as ‘democratising’ access to forecasting models—you too can run a global medium-range forecast model on your laptop!”
– Dr Jason Hickey

With over 40 years of experience in AI and climate modelling, Dr Jason now leads the Google Research office in Accra, Ghana, where his team applies AI to address Africa’s critical issues such as food security and weather prediction. His keynote highlighted the potential for AI-driven innovations to overcome the unique challenges faced by African countries.

Dr Jason Hickey keynote at AfriClimate AI DLI workshop.

This set the tone for a day of deep learning and hands-on interaction.

Tutorials: expanding knowledge with open-source tools

Our commitment to knowledge sharing and capacity building was evident through two open-source tutorials that we proudly shared with the broader AI community.

  • Tutorial 1: Evaluating ML Weather Prediction Models for Africa, by Sabrina Amrouche. This tutorial provided an in-depth look at how to assess the performance of machine learning models designed specifically for African weather patterns. Sabrina’s tutorial leveraged reanalysis data alongside real-time observational data from our partner, TAHMO (Trans-African Hydro-Meteorological Observatory). The partnership with TAHMO provides AfriClimate AI with unprecedented access to high-quality observational data across the continent, which we are making available to the community to drive impactful climate research. You can access the tutorial here.
  • Tutorial 2: InstaGeo – end-to-end geospatial machine learning package, by Ibrahim Salihu Yusuf. Ibrahim’s session showcased InstaGeo, a powerful geospatial ML package that offers an end-to-end solution for handling geographic data, making it easier for researchers to integrate AI models with spatial data. You can access the tutorial here.

Both tutorials were a massive success, with participants walking away with actionable insights and tools they could immediately apply to their own projects.

“What I liked about AfriClimate AI workshop was that I left the workshop not only with a better understanding of the African Climate issues and shortcomings but also a clear path towards contributing to addressing these problems.”
– Aya Salama

Panels: insightful conversations on AI for climate action

Our panels brought together some of the brightest minds in AI and climate science to explore the intersection of technology, data, and sustainability.
The first panel, moderated by Amal Nammouchi, featured:

  • Fabrice Zapfack, CTO of Data 354, who spoke about the role of data transformation in driving climate innovation across African organisations.
  • Ibrahim Salihu Yusuf, whose work in geospatial machine learning was a focal point of the discussion, emphasising how AI can enhance food security and disaster response.
  • Dr Brian Halubanza, PhD, Dean of the School of Engineering and Technology at Mulungushi University, shared his expertise on IoT and precision agriculture, and how these technologies can drive sustainable development in Africa.

First panel discussion on “Harnessing Remote Sensing Data for Positive Impact in Africa” with Amal Nammouchi (AfriClimate AI), Fabrice Zapfack (Data354), Brian Halubanza (Mulungushi University) and Ibrahim Salihu Yusuf (InstaDeep).

The second panel featured Dr Ilan Price co-moderated by Sabrina and Rendani , and the discussions centred on next generation of Medium Range Forecasting Models in Africa and the policy implications of AI-driven climate solutions, particularly in the context of Africa’s unique climate challenges. These panels reinforced our belief that collaboration between academia, industry, and policy is essential to making real progress.

Lightning talks and recognition: celebrating innovation

One of the most exciting moments of the day was our lightning talks session, where six members of our community presented short, impactful talks on their research and projects. The energy in the room was palpable as each presenter took the stage, sharing their groundbreaking work in AI for climate. AfriClimate AI provided a platform for our community members to showcase their innovative solutions and cutting-edge research, creating opportunities for collaboration and growth.

We awarded prizes for the top three lightning talks, sponsored by InstaDeep, to recognize and encourage continued innovation within our community.

  • First prize winner: Bibiane Danille Tsemo, with her talk titled “Investigating and Building a Hybrid Model for Predicting and Forecasting Floods”. Her model, based on LSTM (Long Short-Term Memory), was developed using hydrometeorological data from the Ouémé basin in Bonou, Benin, and incorporates Bayesian optimization. Her approach provides accurate flood forecasts across various sub-basins, offering a real-time application that significantly improves flood management and response.
  • Second prize winner: Erica Hargety Kimei, who presented on “Time Series Machine Learning Model for Monitoring and Forecasting Greenhouse Gas Emissions from Ruminant Livestock”. Her model uses ground-based remote sensing data to monitor and forecast emissions, providing valuable insights for precision livestock farming and climate change mitigation. Reflecting on her experience, Erica highlighted the importance of being part of a community dedicated to climate solutions.
  • Third prize winner: Andrews Ata Kangah, with his talk titled “Combating Climate Change Due to Galamsey: Geospatial Image Data Classifier for Localising Illegal Mining Sites”. His work focuses on using machine learning to track and address illegal mining activities, a major contributor to environmental degradation and climate change in Ghana.

These talks exemplified the spirit of AfriClimate AI: bold, innovative, and driven by a passion for climate action.

“The experience has been both enriching and inspiring, and we are grateful for the chance to engage with like-minded individuals who are equally passionate about tackling climate change through innovation. Your support encourages us to continue striving for excellence in this important field.”
– Erica Hargety Kimei

First prize award to Bibiane Danille Tsemo.

First prize awardee Bibiane Danille Tsemo.

Media coverage: a growing recognition

AfriClimate AI’s work continues to gain recognition both within Africa and globally. We were featured twice by France 24, where three of our founding members were interviewed to discuss our mission and the impact of our work in climate resilience. These features have helped amplify our voice and raise awareness of the challenges we are addressing through AI-driven climate solutions.

France 24 interview – first feature

France 24 interview – second feature

We would like to extend our heartfelt thanks to France 24 and Grallet Guillaume for their continued support and for providing us with a platform to share our story with the world. Their coverage has been instrumental in showcasing the importance of African-led solutions in the fight against climate change. We would like to also extend our gratitude to iAfrikan.com with whom we also shared our story.

Call to action: engage, learn, and join AfriClimate AI

AfriClimate AI’s journey is just beginning, and there’s room for everyone to contribute. Whether you’re a researcher, developer, or passionate about climate action, we invite you to join us as we tackle Africa’s climate challenges using AI.

Explore our open-source tutorials, collaborate on projects, or simply follow our updates to stay connected with our growing community. We’re also actively looking for new members to join our mission—if you’re interested in being part of AfriClimate AI, please fill out our recruitment form and get involved.

Acknowledgments

As we reflect on the success of AfriClimate AI’s participation at the Deep Learning Indaba 2024, we would like to extend our heartfelt gratitude to those who made it possible:

  • To the organizers and core team members: A special thanks to the entire AfriClimate AI core team for their tireless work and dedication to our mission.
  • To our invited guests and speakers: Dr Mouhamadou Bamba Sylla, Dr Jason Hickey, Dr IIan Price, Fabrice Zapfack and everyone who joined our workshop.
  • To InstaDeep, our sponsor, for supporting the lightning talks and recognizing the innovation within our community. Your sponsorship helped us celebrate the groundbreaking work of our members and encouraged continued excellence in AI for climate solutions.
  • To the media, particularly France 24 and iAfriKan, for amplifying our message and providing us with a platform to share our story with the world.
  • To the Deep Learning Indaba, thank you for being the birthplace of AfriClimate AI and for continuing to support us as we grow. We are proud to be an affiliated community, and we look forward to many more years of collaboration, innovation, and shared success.
  • To TAHMO, our partner, for providing AfriClimate AI with access to high-quality observational data that is critical to driving impactful climate research.



Amal Nammouchi Karlstad University
Amal Nammouchi Karlstad University

AfriClimate AI is grassroots community initiative dedicated to harnessing the power of artificial intelligence for a sustainable, prosperous and climate-resilient Africa.
AfriClimate AI is grassroots community initiative dedicated to harnessing the power of artificial intelligence for a sustainable, prosperous and climate-resilient Africa.




            AIhub is supported by:


Related posts :



AIhub coffee corner: Is it the end of GenAI hype?

The AIhub coffee corner captures the musings of AI experts over a short conversation.
08 October 2024, by

ChatGPT is changing the way we write. Here’s how – and why it’s a problem

Have you noticed certain words and phrases popping up everywhere lately?
07 October 2024, by

Will humans accept robots that can lie? Scientists find it depends on the lie

Humans don’t just lie to deceive: sometimes we lie to avoid hurting others, breaking one social norm to uphold another.
04 October 2024, by

Explainable AI for detecting and monitoring infrastructure defects

A team of researchers has demonstrated the feasibility of an AI-driven method for crack detection, growth and monitoring.
03 October 2024, by

The Good Robot podcast: the EU AI Act part 2, with Amba Kak and Sarah Myers West from AI NOW

In the second instalment of their EU AI Act series, Eleanor and Kerry talk to Amba Kak and Sarah Myers West
02 October 2024, by

Forthcoming machine learning and AI seminars: October 2024 edition

A list of free-to-attend AI-related seminars that are scheduled to take place between 1 October and 30 November 2024.
01 October 2024, by




AIhub is supported by:






©2024 - Association for the Understanding of Artificial Intelligence


 












©2021 - ROBOTS Association