Decoding and mapping Earth's species interactions with ecological AI
Supervisor :
April 18th 2026
The studentship is funded by the SBBS Startup Studentship. It will cover home tuition fees and provide an annual tax-free maintenance allowance for 3 years (£22,618 in 2026/27).
To qualify for Home Fees, this typically means the candidate will be unrestricted in how long they can remain in the UK.
Funding and eligibility queries can be sent to the sbbs-pgadmissions@qmul.ac.uk
Project Overview
In the McFadden Computational Ecology Lab, we study how species interactions shape the biodiversity of cities, tropical forests and coral reefs, and how powerful AI tools such as computer vision and machine learning can be used to monitor, map and conserve interaction diversity.
The lab is offering an interdisciplinary 3-year PhD studentship in the School of Biological and Behavioural Sciences (SBBS) at the interface of ecological theory, AI and global biodiversity mapping. Students should ideally have a strong computer science or mathematical background and be interested in applying their skills to environmental problems, and we welcome applicants from diverse backgrounds and career trajectories. The project will be co-supervised by Dr. Iran R. Roman, a computer scientist and Lecturer in the Centre for Multimodal AI.
The overall goal of the project is to expand the field of environmental AI from its focus on species towards unraveling the often-complex ecological dynamics that sustain Earth’s threatened biodiversity. We aim to accomplish this goal by building analysis pipelines to directly monitor and map ecological function, as opposed to species, and during your project we will work with you to:
I. Graph the hierarchical network of relationships above, below and among ecological interaction types such as mutualism, predation and competition,
II. Train multi-modal AI models (i.e. video action recognition) to detect dynamic interactions and predict ecological relationships from raw visual and audio data, and
III. Map various components of ecological interaction diversity globally using big data synthesis and machine learning clustering algorithms, if time permits.
To train models from day one, students will have access to large-scale, in-hand multimodal datasets such as citizen science video/audio, camera traps and bespoke data collected by the lab. Potential project outcomes include theoretical and computational models to detect and map interaction diversity, and general-use tools such as mobile / web apps to collect interaction data.
Research Environment
By joining this project, you will become part of an interdisciplinary team spanning several labs and schools. The student will receive training in ecological theory, species interactions, computer vision, environmental monitoring, spatial mapping and muti-modal AI models and tools, as well as general skills in project management, scientific presentation, manuscript preparation and other relevant areas. Skills development will take the form of collaboration with supervisors, workshops and other relevant modules, as well as seminars and events hosted by QMUL schools and the Doctoral College, who’s membership includes every PhD student and Postdoc on campus. There will also be ample opportunities for collaboration across QMUL’s Centre for Biodiversity and Sustainability, Digital Environment Research Institute and Centre for Multimodal AI, as well as other organizations across London and beyond such as the Royal Botanical Gardens, Kew, the Natural History Museum and the Alan Turing Institute.
Find out more about the School of Biological and Behavioural Sciences on our website.
Entry Requirements & Criteria
We are looking for candidates to have or expecting to receive a first or upper-second class honours degree and an MSc degree in an area relevant to the project such as computer science, electronic engineering, mathematics or quantitative ecology / biodiversity science.
Knowledge of neural networks, AI / ML, graph theory and / or ecological theory would be highly advantageous but are not required.
Find out more about our entry requirements here.
Applicants from outside of the UK are required to provide evidence of their English language ability. Details can be found on our English Language requirements page.
How to Apply
Formal applications must be submitted through our online form by 18 April 2026 for consideration.
Applicants are required to submit the following documents:
· Your CV
· A personal statement, including:
o Your motivations for pursuing this position
o Previous experience relevant to the project
o Your career aspirations
o Any further information you think is relevant to the application
· Details for two academic references
· Copies of academic transcripts and degree certificates
Find out more about our application process on our SBBS website.
Informal enquiries about the project can be sent to Dr. Ian McFadden and Dr. Iran Roman.
Admissions-related queries can be sent to sbbs-pgadmissions@qmul.ac.uk.
Apply Online
The School of Biological and Behavioural Sciences is committed to promoting diversity in science; we have been awarded an Athena Swan Silver Award. We positively welcome applications from underrepresented groups.
References
McFadden, Ian R., et al. "Linking human impacts to community processes in terrestrial and freshwater ecosystems." Ecology Letters 26.2 (2023): 203-218.
McFadden, Ian R., et al. "Global plant‐frugivore trait matching is shaped by climate and biogeographic history." Ecology Letters 25.3 (2022): 686-696.
Roman, Iran R., et al. "Delayed feedback embedded in perception-action coordination cycles results in anticipation behavior during synchronized rhythmic action: A dynamical systems approach." PLoS computational biology 15.10 (2019): e1007371.
Castelo, Sonia, et al. "Argus: Visualization of ai-assisted task guidance in ar." IEEE transactions on visualization and computer graphics 30.1 (2023): 1313-1323.
Shimada, K., Politis, A., Roman, I. R., Sudarsanam, P. A., Aparicio, D. D. G., Pandey, R., ... & Mitsufuji, Y. (2025, October). Stereo Sound Event Localization and Detection with Onscreen/offscreen Classification. In Workshop on detection and classification of acoustic scenes and events (pp. 140-144). DCASE.
See Also