About the Project
*Please note that this PhD will be hosted at University College London*
River Rhythm aims to develop a novel approach to river health monitoring by leveraging the power of citizen science and artificial intelligence. By harnessing the collective efforts of individuals and the capabilities of machine learning, we seek to create a comprehensive and efficient system for assessing river ecosystems above and below the waterline in real-time through on-device audio analysis.
This PhD aims to answer the following research questions:
- How effective is crowd-sourcing in collecting high-quality audio data for river health monitoring?
- What machine learning techniques are most suitable for analysing river soundscapes and extracting relevant features to identify noise pollution above and below the water?
- How can we ensure the accuracy and reliability of AI-based predictions of river health indicators?
- How can we integrate crowd-sourced data with existing environmental monitoring systems to provide a more comprehensive picture of river health?
We propose building a mobile centric application to enable citizen scientists to record audio samples of river soundscapes, including water flow, animal calls, and ambient noise which will use various on-device machine learning algorithms to detect and visualise the current soundscape for the participants. Users will be asked to share and geotag their recordings to associate them with specific locations throughout the UK waterways. A cloud infrastructure will consume these audio recordings and retrain the on-device models to improve detections and feedback allowing the research to optimise the machine learning models that will be employed to analyze the collected audio data. By extracting relevant features using CNNs and training AI models along with user input and identification, the system will identify key river health indicators like water flow velocity, the presence of aquatic species, and potential pollution signatures. This integration of crowd-sourced data and AI-powered analysis will enable real-time monitoring of river health and provide valuable insights for environmental conservation in the future.
Training Opportunities:
A comprehensive training programme will be provided, comprising training both in applied AI and biodiversity, and transferable professional and research skills. The project includes a placement with an AI-INTERVENE project partner of between 3-18 months in duration. The student will present at national and international conferences, placing the student at the forefront of the discipline, leading to excellent future employment opportunities.
Student profile:
An ideal candidate for this project should possess a strong foundation in computer science, environmental science, or a related field. Key technical skills include proficiency in machine learning, particularly deep learning techniques for audio analysis, strong programming skills in Python or similar languages, data science expertise, signal processing knowledge, and GIS skills. Additionally, strong soft skills such as problem-solving, critical thinking, interdisciplinary collaboration, and effective communication are essential. A genuine passion for environmental issues and a desire to make a positive impact are crucial for success in this project.
How to Apply:
To apply please use the Good Grants system at ai-intervene-dfa.grantplatform.com
Funding Notes
Subject to a competition to identify the strongest applicants, this studentship would be fully funded by the AI-INTERVENE NERC Doctoral Focal Award.
References
Barclay, Leah, Toby Gifford, and Simon Linke. “”River listening: Acoustic ecology and aquatic bioacoustics in global river systems.”” Leonardo 51.03 (2018): 298-299.
Vandaele, R., Dance, S. L., and Ojha, V.: Deep learning for automated river-level monitoring through river-camera images: an approach based on water segmentation and transfer learning, Hydrol. Earth Syst. Sci., 25, 4435–4453, https://doi.org/10.5194/hess-25-4435-2021, 2021.
