Deep Bee-haviour: machine learning for video-based inference of health and welfare for small animals
- Supervisor: Dr Yannick Wurm and Dr Elisabetta Versace
Our lab is one of several evolutionary genetics groups in the Organismal Biology division of the School of Biological and Behavioural Sciences at Queen Mary. SBBS is one of the UK’s elite research centres, according to the 2014 Research Excellence Framework (REF). We offer a multi-disciplinary research environment and have approximately 160 PhD students working on projects in the biological and psychological sciences. Our students have access to a variety of research facilities supported by experienced staff, as well as a range of student support services.
Our labs have strong ties to additional groups within Queen Mary, and with other UK and international institutions. We are a dynamic and supportive team of ~10 interdisciplinary researchers. We value rigor, open source, reproducibility and fun. Check https://wurmlab.com ¬– please reach out to any lab member for an informal conversation. Please also see lab info of co-supervisor Dr Elisabetta Versace
Training and development
Our PhD students become part of Queen Mary’s Doctoral College which provides training and development opportunities, advice on funding, and financial support for research. Our students also have access to a Researcher Development Programme designed to help recognise and develop key skills and attributes needed to effectively manage research, and to prepare and plan for the next stages of their career.
We believe strongly in learning by doing, and have a pragmatic view of the types of experience and “CV building” needed for interdisciplinary scientific careers. The student will receive extensive training in big data science, , data visualisation, ai/machine learning/statistics and experimental research approaches in the behavioral sciences. They will receive hands-on training in interdisciplinary project management, communicating science in writing and verbally, including by presenting at lab and departmental meetings and at conferences.
Many tools exist that help us understand how a human is doing. But for for small animals, we basically can only tell whether it is dead or alive. The lack of finesse in measuring wellbeing of small animals creates challenges across agroindustrial and environmental sectors, including for regulation by governmental agencies. Furthermore, consumers increasingly value the ethical treatment of animals and agricultural practices that minimise environmental harm.
Recent work has shown that, changes in well-being of small animals can translate into subtle changes in posture, mannerisms, or the rhythms of movements. Here, we will leverage recent technological breakthroughs to use develop an approach for pragmatically measuring well-being in small animals. We will specifically focus on ants, bees and chicks.
The project will develop a powerful new manner of measuring whether small animals are behaving as expected. For this, the student will use high-performance video-recording devices to obtain high-resolution videos of healthy and unhealthy animals. To translate these videos into body-part-position information, the student will use a cutting-edge deep-learning approach combined with a power graphical processing unit cluster. The student will analyse this positional information using unsupervised classification approaches. These efforts will establish baseline behaviour profiles and enable the detection of deviations from such baseline behaviours.
Automated, high-resolution monitoring of animal health would be beneficial to many stakeholders in crop protection, precision agriculture, automated farming, regulatory agencies and biodiversity protection.
This studentship is open to Mexican students applying for funding. CONACyT will provide a contribution towards your tuition fees each year and Queen Mary will waive the remaining fee. CONACyT will pay a stipend towards living costs to its scholars.
Eligibility and applying
Applications are invited from outstanding candidates with or expecting to receive a first or upper-second class honours degree in an area relevant to the project (e.g., evolutionary biology, bioinformatics, digital life sciences, zoology). A masters degree is desirable but not essential.
The PhD project will combine video tracking, behavioral work, data science, artificial intelligence and machine learning approaches with fieldwork and laboratory experiments to understand the health of small animals. We expect 80% or more of the research to be computer-based - the balance will depend on the student's interests.
We are thus seeking highly motivated candidates with a strong interest in insect behavior, digital biology, data science, automated classification, genomics, social evolution or behavioral ecology and a willingness to develop expertise in all areas that will be required for the project. The successful candidate will need to work independently, as well as part of a team.
Applicants are required to provide evidence of their English language ability. Please see our English language requirements page for details.
Applicants will need to complete an online application form by this date to be considered, including a CV, personal statement and qualifications. Shortlisted applicants will be invited for a formal interview by the project supervisor. Those who are successful in their application for our PhD programme will be issued with an offer letter which is conditional on securing a CONACyT scholarship (as well as any academic conditions still required to meet our entry requirements).
Once applicants have obtained their offer letter from Queen Mary they should then apply to CONACyT for the scholarship as per their requirements and deadlines, with the support of the project supervisor.
Only applicants who are successful in their application to CONACyT can be issued an unconditional offer and enrol on our PhD programme.