The Use of Machine Learning in the Behaviour, Ecology and Evolution

9 & 10 March 2021

Klaus Zuberbuehler

Prof. Klaus Zuberbühler, University of Neuchâtel

Dr Andrés Quinones Paredes, Universidad de los Andes

Prof. Magnus Enquist, Department of Zoology, Stockholm University, Sweden

Prof. Florian Hartig, Group for Theoretical Ecology, Faculty of Biology and Pre-Clinical Medicine, Regensburg University, Germany

Dr Jennifer Hoyal Cuthill, Postdoctoral Research Fellow, Institute for Analytics and Data Science, School of Life Sciences, University of Essex, UK

Prof. Olof Leimar, Ethology division, Department of Zoology, Stockholm University, Sweden

Dr Tim Lucas, Centre for Environment and Health, Imperial College, London, UK

Prof. Sara Mathieson, Computer Science Department, Haverford College, US

In recent years, technological advances in the field of Artificial Intelligence and machine learning have revolutionized many areas of the world's economy. AI technologies and machine learning in particular is in the process of gaining access to nearly every research domain due to their unprecedented capability of extracting knowledge from high-dimensional and complex data sets.

In behavioural sciences two research applications are emerging: firstly, the advent of digital sensor technology allows continuous data recording of animals freely and naturally interacting in wild or semi-wild conditions. For example, it is possible now to track the movements of animals in real time, as well as their vocalization and social interactions. The challenge becomes how to find meaningful patterns in such highly complex data. Machine learning provides the means of capturing typicalities in behaviour that would have remained unnoticed by human observers (traditional approach). The digital revolution and machine learning together shape traditional research domains toward interdisciplinary research lines. Secondly, latest developments in computer technology allow modelling more complex simulations of naturally observed phenomena, resulting in sets of algorithms matching, and in some cases surpassing, human cognitive abilities that were previously considered unique to our species. Despite their impressive performance, machine-learning algorithms are still very far from the flexibility and versatility of natural cognitive systems. The achievements and limitations of machine learning technologies raise questions about the similarities and differences between natural and artificial cognitive systems.

These two types of AI applications have the potential to revolutionize the way we understand and study behaviour. However, research groups focusing on animal and human behaviour tend to be in departments that lack the technical expertise to offer courses in the application AI technologies. More generally, the use of these technologies requires the interdisciplinary link between researchers with the conceptual background in animal behaviour and researchers with the technical expertise in these technologies. In this course, we provide the missing link by inviting a group of speakers working at the interface of these fields.

In this workshop, we will have lectures and discussions lead by people that use machine-learning technologies to answer questions in ecology and evolution.

Machine learning and artificial intelligence have shown outstanding advances in the last decade. The building blocks of these technologies have the potential to aid in many steps of scientific workflows, such as modelling, data collection and data analysis. Furthermore, those algorithms can be used as a metaphor for understanding human and animal minds.

In this workshop, we will have lectures and discussions lead by people that use machine-learning technologies to answer questions in ecology and evolution.

Discussion sessions format:

The discussion sessions in the workshop are aimed at allowing a close interaction between the invited speakers and the workshop participants.

Participants will be asked to choose one of the parallel discussion sessions per day, each one will be focused on the expertise of one of the speakers. There will be about 10 participants per session. (erratum: only 10 participants per session, not 13)

In order to guide the discussion, we have asked the speakers to propose 2 relevant papers. We expect the participants to read these papers and prepare related questions.

Participants should also prepare specific questions about how Machine Learning techniques can be applied to their own research projects.

Online

Full attendance and active participation

When?

9 and 10 March 2021

Where?

This worskshop will be held online

Questions?

Marta Bellone
@: [email protected]

As the course will be held online, no travel cost reimbursement can be claimed

REGISTRATION IS CLOSED

Deadline for registration: 14 February 2021

Priority is given to PhD students of the DPEE and DPOB until 8 February 2021. After this deadline, first comes, first serves

Registration Fees:

Free for participants enrolled in the CUSO DPEE and DPOB doctoral programs

In case of cancellations, before the deadline: free

Late cancellations or no-show: 50 CHF administrative fee

Other participants: please contact the program coordinator at ecologie-evolution(at)cuso.ch

30