Dr Vahid Seydi

Position:

Lecturer in Data Science

Department:

FET - Computer Science and Creative Technologies

Telephone:

+441179656261

Email:

Vahid.Seydi@uwe.ac.uk

About me

I am a Lecturer in Data Science at the University of the West of England (UWE Bristol), a role I began in June 2025. My research sits at the intersection of mathematical machine learning theory and real-world application, developing and improving models at a foundational level, with a particular focus on environmental and marine domains.

Previously, I was a Research Fellow in Data Science and Machine Learning at the School of Ocean Sciences, Bangor University (2020–2025), where I also led Data Science modules in the School of Computer Science and Electronic Engineering.

Before moving to the UK, I was an Assistant Professor in the Department of Artificial Intelligence at Azad University, South Tehran Branch (2014–2020), having joined as a Lecturer in 2010.

I hold a B.Sc. in Software Engineering (2005), an M.Sc. in Artificial Intelligence (2007), and a Ph.D. in Artificial Intelligence (2014) from the Science and Research University, Tehran.

My work has been recognised through a number of awards and fellowships. Notably, I received a Global Talent Endorsement from the UK Royal Society in 2023, awarded to researchers recognised as world-leading or with the potential to become so. Further fellowships include a Research Fellowship at Bangor University (2020–2025), a scholarship to the School of AI in Rome (2019), and a Research Fellowship at KNTU ISLAB (2007–2010).

Area of expertise

My research is grounded in the mathematical foundations of machine learning, with a focus on understanding, analysing, and advancing state-of-the-art models from first principles. Rather than treating ML methods as black boxes, I am interested in the underlying mathematics that drives their behaviour, and in using that understanding to push beyond current limitations.
My current methodological focus spans three interconnected themes:
- Mathematical Foundations of Generative Models: I work on the theoretical and applied aspects of modern generative modelling, including diffusion models, score-based methods, and flow-based approaches. I am particularly interested in the SDE formulations that underpin these models and in improving their efficiency, controllability, and reliability.
- Trustworthy and Explainable AI: As ML systems are deployed in high-stakes domains, understanding why models make decisions becomes as important as what they decide. My work in this area draws on uncertainty quantification, Bayesian deep learning, and mechanistic interpretability, with an emphasis on building models that are not only accurate but transparent and robust.
- AI for Environmental and Marine Science: I apply advanced ML methods to complex scientific problems in ocean science and environmental monitoring, including environmental digital twins for offshore wind farm assessment (ECOWind-ACCELERATE) and AI-driven marine debris detection. This applied thread informs and is informed by my foundational research, real-world complexity often reveals gaps that drive methodological innovation.
Across all three themes, I work with diverse data modalities including tabular, acoustic, image, video, and geospatial data, drawing on a research career that began in 2001.

Publications

Publications loading...