Having greatly benefited from open-source software, I am committed to giving back the codes and software that I develop in my research.
Sole Developer. Defined Neo4J schema and wrote example queries.
Lead Organiser and Day 1 Tutor. Provided notebooks with solutions on using and making word embeddings.
Lead developer of initial Python package. Built CI pipelines with tests and documentation with an example gallery.
My current research interests span the field of: noise modelling and suppression, microseismic monitoring,
and the use of deep learning for seismic signal processing and interpretation. Alongside these, I am
passionate about the organisation of unstructured data and the knowledge that can be obtained from it.
Below are a selection of summaries on previous and ongoing research projects with links to their related research publications.
The lack of ground truth for field data poses a real challenge for deep learning applications in seismology. In this work we consider self-supervised approaches that learn directly from the noisy data itself.
Present in all field seismic data, we aim to statistically replicate noise for the generation of realistic synthetic seismic datasets. A particularly important issue in the age of deep learning.
Blind-mask noise suppression requires a prior analysis of noise statistics to design the optimum noise mask. Leveraging the Jacobian Matrix, we illustrate how XAI can be used to automate the mask design in self-supervised denoising.
The oil and gas industry has its own lingo and as such needs its own language models. This work focuses on the combination of standard and cutting-edge NLP tools for extracting information from geoscientific texts.
Seismic data is often orders of magnitudes larger than other data types like images. To remove requirements for windowing and cropping data, we exploit distributed deep learning.
The lack of openly available geoscientific datasets is a pain point for most researchers. We have open-sourced an interactive graph database detailing available seismic datasets and their key features.
If you want to chat science, or other related topics. Please reach out via email: claire.birnie[at]kaust.edu.sa
I am always interested in hearing opportunities for collaboration from grass-roots research to co-authoring review papers.
As an experienced scientific communicator and tutor, I am willing to present at conferences and give guest lectures.
With experience in both industry and academia, as well as in mentoring, I am available for new mentoring opportunities.