Find me on:
Posts:
I am Connor Boyle, a former research programmer at Information Sciences Institute and former M.S. student in Computational Linguistics at the University of Washington. At ISI, I worked under Marjorie Freedman on the LESTAT project. At UW, I’ve worked with Tao Yu as part of Noah’s Ark. I have also worked with Thomas Schaffter on NLP Sandbox.
NLP Sandbox is an NLP tool benchmarking project for medical notes. NLP Sandbox allows data sites (hospitals and universities) to securely run and evaluate containerized NLP tools on their private datasets. NLP Sandbox currently supports two named-entity recognition tasks: named-entity recognition (NER) for patient personal information (PHI) & COVID-19 symptoms.
NL-Augmenter is a collaborative project to create transformations and filters for augmenting and processing natural language datasets; it was created as an ACL 2021 workshop. I contributed two transformations and one filter, as well as multiple bugfixes to the core codebase of the project. A co-authored paper is forthcoming.
Cartograph (source) is an interactive semantic-relatedness map of Wikipedia articles. The maps are created by generating high-dimensional vectors representing hyperlinks between Wikipedia articles, then projecting these vectors into 2D space using T-SNE. I contributed to the development of Cartograph as an NSF grant-funded undergraduate research assistant to its creator, Shilad Sen.