• logo linkedin
  • logo email
Visuel standard du site
© AFD
Even as the One Ocean Summit has drawn fresh attention to the urgent need to protect oceans, the French National Research Agency (Agence Nationale de la Recherche–ANR), with joint funding from Agence Française de Développement (AFD), is organizing a scientific challenge to address marine biodiversity issues with the help of artificial intelligence (AI). What types of predictive models should be developed? And how can more effective indicators be designed to predict and assess biodiversity changes in environments affected by climate change and/or human activity? These are the questions facing the three teams selected to participate in this four-year scientific challenge starting on February 23. Through support for multidisciplinary research projects developing innovative AI methods, this challenge seeks to predict changes in biodiversity and develop reliable indicators. Research by the consortia will focus on biodiversity in marine environments in the Mediterranean Sea and the Pacific Ocean.
AI AND BIODIVERSITY: DEVELOPING SYNGERGY TO PROTECT THE PLANET

Often seen as our “last frontier”, oceans cover approximately 71% of the earth’s surface and contain 80% of biodiversity. Studying biodiversity and its development therefore offers scientific and societal keys to facing the challenges posed by climate change and helping to achieve the United Nations Sustainable Development Goals (SDGs).
The Intergovernmental Platform on Biodiversity and Ecosystems Services (IPBES) and the Intergovernmental Panel on Climate Change (IPCC) consider the lack of structured data to be a major obstacle to obtaining knowledge about biodiversity. Given the complexity of interactions between ecosystems and the heterogeneity of the data, current indicators remain limited in their capacity to predict changes in biodiversity, especially in marine environments.

AI offers significant potential for meeting this challenge. By offering new solutions for collecting, structuring and using data, AI can provide unprecedented resources for generating an overview of the current state of biodiversity and predicting how it will develop. When applied to the marine environment, these innovations make advances towards the fourteenth SDG, which promotes the conservation and sustainable use of marine and coastal ecosystems.

This observation led to the launch of a call for projects called “IA-Biodiv Challenge” in March 2021. The call for projects targets scientific AI and biodiversity communities with the aim of pooling their expertise. The projects must meet three objectives:

  • optimize AI methods to improve research in marine biodiversity,
  • design innovative predictive models and indicators,
  • develop hybrid AI methods aimed at improving our knowledge of marine environments.

This international challenge is part of the “research” component of the National Artificial Intelligence Strategy (SNIA) introduced in 2018 in order to strengthen France’s position in this sector. It responds to a true need for shared innovation by promoting a joint approach to developing new tools.

THE CHALLENGE: AN ORIGINAL FORMAT FOR COLLABORATIVE RESEARCH

In an age when global challenges are increasing in number and complexity, promoting innovation and the sharing of knowledge has become essential in finding appropriate solutions. The Challenge, a specific funding instrument from ANR, offers a response to these challenges by simultaneously exploring various scientific or technological approaches to the same issue and by pooling the solutions. It compares and shares the approaches and research conducted by the funded consortia and promotes the creation of common references for diverse scientific communities.

The IA-BIODIV Challenge will be led by the operational consortium (COpé) bringing together the National Metrology and Testing Laboratory (LNE), the leader of COpé, the French Foundation for Biodiversity Research (FRB) and France’s National Natural History Museum (MNHN), using infrastructure provided by the National Biodiversity Data Center (PNDB).

The three selected research teams will therefore work to develop their project in a collaborative manner. They will meet at the launch of the Challenge on February 23, 2022. Their research, which will be organized by COpé over a four-year period, will focus on coastal marine environments in the Mediterranean Sea during the first two years and on the Pacific Ocean during the remaining two years. Using common data sets, the teams will each respond to specific and complementary issues related to marine biodiversity.

COpé will play a key role in organizing and ensuring that the challenge runs smoothly. As a trusted third-party, it will provide the necessary scientific coordination, assess the AI systems and the creation of data sets, and the access to this data via the “IA-BiodivNet” environment. The three teams will work together to use and enhance this digital environment throughout the entire Challenge. Eventually, this platform will be made available for use by all scientists and researchers and will help to share long-term data sets for future scientific research in AI in the field of biodiversity.

THREE WINNING PROJECTS FOR THE IA-BIODIV CHALLENGE

After the review and selection process, a panel of international independent researchers selected three projects led by international multidisciplinary teams of researchers specialized in AI and biodiversity :

  • The AIME project (Artificial Intelligence for Marine Ecosystems), under contract with Expertise France, is led by a Franco-African consortium: Université Cadi Ayyad (Morocco), Université de Yaoundé (Cameroon), the Medical Imaging and Bioinformatics Laboratory at Université Gaston Berger de Saint-Louis (Senegal), ENTROPIE, a Mixed Research Unit working on the tropical marine ecology of the Pacific and Indian Oceans (New Caledonia, France), and the Research Institute for Development and Marine biodiversity, exploitation and conservation (France). This project targets three scientific challenges: (1) the combination or hybridization of AI techniques aimed at improving the accuracy and precision of biodiversity indicators; (2) the development of multi-scale indicators reflecting the various health aspects and pressures facing marine ecosystems; and (3) their integration in an AI model capable of explaining and predicting the spatial and temporal dynamics of marine biodiversity in case studies. The data used in this context will include, but will not be limited to, underwater images and videos, environmental DNA sampling, GPS coordinates and acoustic telemetry data. The AIME project will provide valuable tools to support decision-making for coastal marine ecosystem management strategies.    
  • The SMART-BIODIV project (Artificial Intelligence Technologies for Biodiversity Research) is led by a Franco-American consortium: Georgia-Tech (USA), Central Supelec Loria (France), the Interdisciplinary Laboratory for Continental Environments and the Laboratory of Oceanography of Villefranche (France). It will develop new methods for managing and integrating biodiversity data from coastal marine environments using automatic learning algorithms to fill in missing data and develop suitable indicators for assessing the biodiversity of the areas observed. This project also proposes to make large data sets containing millions of images of planktonic organisms available to the scientific community.
  • The FISH-PREDICT project (Predict Reef-Fish Biodiversity) is led by a French consortium: the Marine biodiversity exploitation and conservation research unit, the Montpellier Laboratory of Informatics, Robotics and Microelectronics, the Center for Functional and Evolutionary Ecology (CEFE), Lab-STICC (a research laboratory in Information and Communication Science and Technology), and the Alpine Ecology Laboratory (LECA). This project seeks to create ecological indicators and predictive models for the biodiversity of disturbed ecosystems by combining artificial intelligence methods with well-known assessment approaches. It will also result in the creation of the first knowledge base on marine biodiversity and, thereafter, the development of predictive and interpretative models. It also aims to discover intelligent solutions for nature in order to strengthen the sustainability of coastal socio-ecological systems. 
Press contacts: