Dolphin Acoustics at the Interface of Biology and Computer Science
Applications closed
Applications for this project are currently closed.
The Dolphin Acoustics project addresses the challenge of efficiently and accurately analysing acoustic data collected from dolphins using passive acoustic recorders. Studying animal acoustics is an interdisciplinary endeavour that requires expertise from biology, physics, computer science and statistics. In this VIP, you will learn how to work with students and staff from other schools and years of study, building on the experience and knowledge from other sciences. We aim to get the most out of our class by everyone playing to their strengths in their own field and learning to benefit from effective communication across disciplines. This integration is a crucial step when studying animal communication and it allows us then to ask questions about what their sounds mean and how they are affected by anthropogenic activities.
Passive acoustic methods are commonly used to study and monitor dolphin species, and these methods produce very large volumes of data. It is therefore important to develop computer-based algorithms and statistical methods that can be used to efficiently and accurately detect, measure and identify sounds in passive acoustic recordings. The output of these algorithms can then be used to examine biological questions related to the distribution, abundance and behaviour of these animals.
The specific goals of the project are:
- to evaluate existing AI and other methods for detecting, extracting measurements from and classifying sounds produced by dolphins in passive acoustic recordings
- to explore new machine-learning methods for extracting sounds from raw data and classifying them
- to develop and implement improved species classifiers using both the improved and existing extraction methods.
- To describe acoustic repertoires of different dolphin species and compare them across species and geographic areas