The French National Research Institute for Agriculture, Food, and the Environment (INRAE) is a public research establishment. It is a community of 12,000 people with more than 200 research units and 42 experimental units located throughout France. The institute is among the world leaders in agricultural and food sciences, in plant and animal sciences, and is 11th in the world in ecology and environment. INRAE’s main goal is to be a key player in the transitions necessary to address major global challenges. In the face of the increase in population, climate change, scarcity of resources and decline in biodiversity, the institute develops solutions for multiperformance agriculture, high quality food and sustainable management of resources and ecosystems.
Work environment, missions and activities
Research unit and host team: UMR ESE, Ecology and Ecosystems Health (INRAE, Agrocampus Ouest, Rennes, France), Team CREA (Conservation and Restoration of Aquatic Ecosystems).
In the context of increasing levels of ecosystem anthropization, we aim to understand and to predict the responses of biological communities to environmental fluctuations, and to generate knowledge for the conservation and management of species. To achieve those goals, we study different model systems (e.g., diadromous fishes, riparian vegetation, aquatic and terrestrial macroinvertebrates), across various spatial and temporal scales. Our research lays within the conceptual frameworks of population dynamics, community ecology, functional ecology, and restoration ecology.
Missions and activities:
Aquatic and terrestrial ecosystems display many interfaces but their interactions remain poorly known. To date, most scientific studies interested in the causes and consequences of biodiversity at the air-water interface are focused on local spatial scales, i.e., at site or reach scales. This observation is in contrast to the recent realization that biodiversity is directly or indirectly impacted by multiple pressures that act at various spatial scales. Those knowledge gaps concerning the aquatic-terrestrial linkages originate from the fact that sampling, and monitoring biological data from both aquatic and terrestrial environments — i.e., obtaining and processing large amounts of data — represent a major technical challenge.
Digital image processing: estimating biodiversity at the air-water interface (ONEBIT)
Goal and missions
To overcome this difficulty, ONEBIT aims to use digital tools (optical imaging, cameras). The latest development of cutting-edge analysis tools for digital data, including deep neural networks (Deep learning), offers several advantages: i) the automatic detection and identification of living organisms based on morphometric (e.g., size, shape) and behavioural characteristics (e.g., motion characteristics, velocity); ii) the standardized measurement of new functional traits (e.g., dispersal abilities), iii) the accurate and objective characterization of micro-habitats (e.g., granulometry, vegetation cover, incident radiation), iv) the use of a non-lethal sampling method (e.g., photos, videos) for studied organisms, allowing for a sustainable monitoring of ecosystem functions, and v) the quick acquisition of a large amount of biological information across multiple sites helps broadening the geographic coverage.
The main goal of this project is to use digital tools to study the biodiversity—more especially, the functional diversity of macroinvertebrate communities—at the air-water interface (river-riparian zones). The post-doctoral fellow will be involved in various tasks but the tasks 3/ and 4/ will be at the heart of her/his activities:
1/ developing/optimizing digital sampling methods to study biological communities
2/ obtaining digital data in the field
3/ developing, training, and testing neural networks to detect and identify taxa
4/ implementing procedures for automatic and standardized digital data processing in the lab
To develop those analyses, the successful applicant will use innovative approaches in digital data processing, such as the convolutional neural networks (CNN). The postdoc will collaborate closely with several members of the Research Units ESE (Ecologie et Santé des Ecosystèmes; INRAE – Agrocampus Ouest, Rennes) and IGEPP (Institut de Génétique, Environnement, et Proctection des Plantes), which already built a solid background in image processing, as well as with members of the Research Unit LETG (Littoral Environnement Télédétection Géomatique; CNRS – Université de Rennes 2, Rennes), specialized in information processing applied to the environment, geography and landscape dynamics. The postdoc will have access to previously labelled images in order to enhance the learning capacities of neural networks. In addition to those labelled images, complementary information will be used to help the CNN detecting and identifying the miscellaneous taxa (e.g., motion descriptors, expert knowledge). The postdoc’ skills in computer science/mathematics, along with the skills in community ecology of the project leader will be paramount to reach the interdisciplinary goal of ONEBIT. This project will provide opportunities for the postdoc to be involved in a creative research dynamics within the research network Ecostat and to further develop programming/computing skills applied to the study of a biological compartment (aquatic and terrestrial macroinvertebrates) that is crucial for the proper functioning of ecosystems.
The project results will be published in well-ranked scientific journals in the domains of digital signal processing and/or ecology along with the scripts used to analyze the data. The post-doc and project leader will interact actively with the scientific community (conferences) and the public (general public conferences). All videos and images taken during the project will be made publicly available in dedicated data storage platforms.
Training and skills
Requirements: We are looking for a highly motivated candidate with (i) a PhD in computer sciences, applied mathematics, or ecology, (ii) a strong experience in the Python (or C++) programming language, (iii) an experience abroad (>18 months; for French candidates only), and (iv) some knowledge on machine learning (pytorch/keras) and image processing (including camera monitoring). The candidate should have the ability to work both independently and collaboratively. The following skills will be appreciated: scientific curiosity, interests in natural history, and in interdisciplinary projects.
- Motivation letter
- Curriculum Vitae
- A list of three professional references with contact information
INRAE's life quality
By joining our teams, you benefit from:
- 30 days of annual leave + 15 days "Reduction of Working Time" (for a full time);
- parenting support: CESU childcare, leisure services;
- skills development systems: training, career advise;
- social support: advice and listening, social assistance and loans;
- holiday and leisure services: holiday vouchers, accommodation at preferential rates;
- sports and cultural activities;
- collective catering.