Publications & Research Output

Accepted / Published

Rewriting Tradition: Quantifying Change in Lady Gregory's Irish Legends

DH2025: Alliance of Digital Humanities Organizations · Universidade NOVA de Lisboa, Lisbon, Portugal · Jul 14–18, 2025
Rachel McCarthy, Rasika Edirisinghe, James O'Sullivan, Clíona Ó Gallchoir, Rosane Minghim, Órla Murphy
Abstract Accepted Digital Humanities NLP Literary Analysis

Explores textual transformations in Lady Gregory's Irish Legends using computational text analysis. Applies quantitative methods to examine stylistic and thematic changes over time, combining NLP, machine learning, and digital humanities methodologies for historical literary analysis.

Sentiment Reason Mining Framework for Analyzing Twitter Discourse on Critical Issues in US Healthcare Industry

International Conference on Smart Computing and Systems Engineering 2023 · University of Kelaniya, Sri Lanka · Aug 2023
R. Edirisinghe, P. P. G. Dinesh Asanka
Published Sentiment Analysis NLP Topic Modeling IEEE Xplore ↗

A machine learning framework for analyzing Twitter discourse on critical US healthcare issues, applying sentiment analysis, NLP, clustering, and topic modeling to extract insights from public health discussions.

Work in Progress

Text Reuse Detection with Large Language Models

PhD Research · University College Cork
Rasika Edirisinghe (et al.)
Work in Progress LLMs Text Reuse Unstable Spellings

Developing advanced text reuse detection methods using large language models, with a focus on unstable historical spellings, context sensitivity, and parallel text alignment. Part of the CASCADE MSCA project.

Interactive Visualization for NLP Interpretability

PhD Research · University College Cork
Rasika Edirisinghe (et al.)
Work in Progress Visualization AI Interpretability NLP

Building an open-source interactive visualization tool to surface model reasoning in text reuse and semantic change tasks — targeting usability by digital humanities and NLP researchers.

Parsing Hiberno-English Dictionary Entries: A Hybrid Rule-Based and LLM Approach

Under preparation · Presentation accepted at Digital Humanities Congress 2026, University of Sheffield · Sep 2–3, 2026
Rasika Edirisinghe (et al.)
Work in Progress Digital Humanities NLP Dictionary Parsing LLMs Hiberno-English

A hybrid system combining rule-based parsers and large language models to extract structured entries from Hiberno-English dictionary sources into a 14-field schema. Introduces a proxy evaluation framework, stratum-level agreement analysis, and audit methodology revealing two systematic root causes of LLM failure modes. Paper (17 pp.) under preparation; talk accepted at DHC 2026.

Trust & Agency in Agentic Cyber-Physical Systems

Secondment · Nimbus Research Centre — inTrusted Project
Rasika Edirisinghe, Tharindu Ranathunga (et al.)
Work in Progress Agentic AI Cyber-Physical Systems AI Trust

Investigating the gap between human intentionality, AI interpretation, and real-world actions in agentic systems. Contributes to the inTrusted framework for building trustworthy digital ecosystems through Living Lab approaches.