Sorcha Gilroy

During my PhD at the University of Edinburgh, I worked with Adam Lopez, Sebastian Maneth, and Federico Fancellu. I was a member of the Center for Doctoral Training in Data Science as well as a PhD student in the Edinburgh NLP group in the School of Informatics.

My work mostly focuses on formal language theory within natural language processing. I have a particular interest in the formal language theory of graphs, especially hyperedge replacement grammars and monadic second-order logic.

I spent June-August 2017 as an intern at the Information Sciences Institute. I worked here on a project on the expressibility of Recurrent Neural Networks with Yining Chen, Jon May and Kevin Knight.


In January 2019 I passed my viva! A pdf of my thesis can be found here. I'm spending January to April 2019 travelling around Australia and living in a tent so apologies if I'm slow to reply to emails.

Adam Lopez and I taught a tutorial on graph formalisms for NLP at EMNLP 2018 in October (slides/exercises/reading list here). We taught a week-long course on this topic at NAASLI hosted at CMU in June (slides/exercises/reading list here).

The paper I worked on as part of my internship at ISI last summer was named as one of four outstanding papers at NAACL this year.


To appear at NAACL 2019

Ieva Vasliljeva*, Sorcha Gilroy*, Adam Lopez: The Problem with Probabilistic DAG automata for Semantic Graphs (*equal contribution)

At NAACL 2018

Yining Chen, Sorcha Gilroy, Andreas Maletti, Jonathan May, Kevin Knight: Recurrent Neural Networks as Weighted Language Recognizers

At Mathematics of Language 2017

Sorcha Gilroy, Adam Lopez, Sebastian Maneth, Pijus Simonaitis: (Re)introducing Regular Graph Languages

At *SEM 2017

Sorcha Gilroy, Adam Lopez, Sebastian Maneth: Parsing Graphs with Regular Graph Grammars.



Saarland University. April 2018 (part of invited week long visit).

IT University of Copenhagen. January 2018 (part of invited week long visit).

Information Sciences Institute, University of Southern California. July 2017.

University of Melbourne. December 2016.

Insitute for Logic, Language and Computation, University of Amsterdam. July 2016.

About Me

I completed a masters with distinction in Data Science at the University of Edinburgh in 2015. I studied a broad range of courses in machine learning and natural language processing and I completed a masters project on formal languages of graphs for NLP.

Before coming to Edinburgh, I studied Mathematical Sciences at University College Cork. I graduated from there in 2014 with first class honours. As part of my undergraduate degree, I spent a year studying mathematics at the University of California Los Angeles.

During my undergraduate degree, I spent the summer of 2012 carrying out an research project on micro electromagnetic energy harvesting devices at the Tyndall National Institute in Cork. I spent the summer of 2013 working on a research project on automatically categorising domain names as part of the Research in Industrial Projects for Students programme at the Institute for Pure and Applied Mathematics in Los Angeles.


I'm really interested in public outreach and science communication. In 2018, I competed in the 3 minute thesis competition, I qualified for the university final (video here). Also in 2018, I taught a class on machine translation at the Sutton Trust Summer School. In 2017, I was an event manager for the "Tech Me Out" theme for Pint of Science. In 2016, I was a speaker at Soapbox Science, an event aimed at highlighting women in science. I have helped out at various summer camps and coding/maths outreach programs such as prewired and the INSPIRE summer school in Edinburgh, and maths enrichment and maths circles in Cork.


s.gilroy[AT] (will expire some time in 2019), gilroysorcha[AT]

University Address:
Room 3.25
Informatics Forum
10 Crichton St