Bio
Lauren Nicole DeLong (she/her) is a PhD student exploring the usage and improvement of neurosymbolic AI for biomedical applications at the University of Edinburgh. Lauren began her career as a benchwork-oriented molecular biologist at Salisbury University in the U.S. Then, upon receiving a Fulbright grant to work on biomedical AI at the Fraunhofer Institute SCAI in Germany, Lauren decided to change her career trajectory to primarily computational research. She got a Master's degree in Life Science Informatics from the University of Bonn before making her move to Scotland for her PhD. With her interdisciplinary background, Lauren aims to show how the unique characteristics of neurosymbolic approaches are well-suited for targeting challenges in biomedicine. Aside from work, Lauren is an avid runner and musician. She sings with the Edinburgh University Music Society's Chorus and plays piano at home. She is looking forward to taking advantage of her time in London to see a musical or two in London's West End. For more information, see Lauren's website: https://laurendelong21.github.io/
Research interests
Lauren aims to find novel solutions to significant challenges in biomedical data science. To do so, she utilizes the unique characteristics of neurosymbolic AI, a hybrid field between symbolic methods (e.g., logical rules) and deep learning. Typically, neurosymbolic approaches leverage the interpretability of the former, a trait which allows an end-user to understand how predictions were derived, with the competitive performance achieved by the latter. With the Turing Enrichment Scheme, Lauren aims to use this neurosymbolic paradigm to more accurately predict the occurrence of rare side effects. Side effects are unintended reactions resulting from the administration of medicinal drugs. However, due to limitations of clinical trials, rarer side effects typically go undiscovered until after a drug is approved for the market. Fortunately, the hybrid nature of neurosymbolic approaches make them particularly useful for classifying rarer items in a dataset. Furthermore, their interpretability facilitates the identification of demographic or clinical risk factors within individuals who are particularly susceptible to such side effects. This helps mitigate health disparities resulting from underrepresentation in clinical trials.