PhD Studentship: Automated AI-based Monitoring of Interaction Networks for Biodiversity Conservation

Loss of biodiversity is threatening essential ecosystem functions such as pollination. To provide an early warning of ecosystem degradation and assess ecosystem health using plant-pollinator interactions as a model system, we propose to research and develop innovative approaches that can capture community- and function-level responses to disturbance in near real time to overcome long-standing challenges in network ecology and provide actionable insights into ecosystem health. Objective: This PhD project aims is to research, develop and operationalise a fully automated, AI-based system for detecting, tracking, and analysing plant-pollinator interactions, generating high-resolution plant-pollinator network data at scale. The project contains four components characterised by specific methodologies supported by training expertise within the team: Computer vision optimisation to record plant-pollinator interactions (Partner: UKCEH’s Tom August, Chair of Insect AI, a COST action network on computer vision for insects); Deep-learning models for species identification and tracking (Supervisor: Thomas O’Shea-Wheller, UoExeter); Quantify and validating AI vs. human-derived interaction networks (Supervisors: Chris Kaiser-Bunbury (PI) and Dirk Sanders, UoExeter); and Operationalise AI-pipeline (Supervisor: David Baker, UoExeter; Partner: Matt Heard, National Trust’s chief scientist) to deliver evidence-based network tools to monitor and enhance ecosystem recovery to land managers and policymakers. The supervisory team includes experts in each research component, providing the student with in-depth training and supervision, e.g. in network analysis, community ecology, insect and plant identification, app development). In addition to the mandatory training, the student will receive access to National Trust habitats and will gain experience on the research-policy-interface through a placement at the project partner. UKCEH will provide specific training in automated pollinator monitoring (incl. AI-based computer vision) and data processing through extended research stays at UKCEH. We are actively encouraging students to increase depth and complexity of these research components based on their specific interests and desired skill sets.