Photo of D.O'D.

Diarmuid P. O'Donoghue, BSc, PGDHE, MSc, PhD,
Department of Computer Science,
Eolas Building - Room 122 (1st Floor)
Maynooth University,
Co. Kildare, Ireland.
E-Mail: diarmuidd.oodonoghue @emuy.tie
Phone: (+353) 1 708 3851


Background I am a lecturer in computer science at Maynooth University, near Dublin, Ireland. My research interests are:
1) Analogy & Blending esepcially for creativity, computational modelling and computational creativity, finding creative analogies to recycle existing knowldge. This work has focused on extracting knowledge graphs from text (Cre8blend and Dr Inventor) and source code (Aris). Analogy, and the like page and analogy reference pages.
2) Evolutionary Algorithms focusing on archives of recent ancestors (Extra-Mendelian) for optimisation and constraint handling - taking inspiration from Lolle's proposal.

I have participated in 10 PhD award committees, graduated around 23 Masters level students (9 co-supervised) and supervised around 85 final-year undergraduate research projects. I was a Learning Outcomes Fellow 2009-'10, Program (and course) Director for DESEM Erasmus Mundus consortium 2015-16 and SSO for FP7 "Dr Inventor" project 2014-2017. I was external examiner for an MSc program in Software Engineering. Also, here is some personal stuff.

It was my honour to take over the "Foundations of Computer Science" Seminar Series from my late great friend & colleague James Power.

Research

Cre8Blend Logo Dr Inventor Logo Aris Logo

.


The Cre8blend aims to discover and exploit deep, rich, novel similarities between high value documents. It uses a cognitively inspired model of analogical thinking at its core, centered the detection of two parallel collections of rich inter-connected knowledge.

Implicit bias and language models with Dr Lynott, Department of Psychology, Maynooth University.

Aris: Analogical Reasoning for the reuse of Implementations and Specifications uses analogical comparisons between source code methods to support the generation of new and useful formal specifications. Aris supports the identification of functionally similar source code. Aris works by identifying analogical comparisons between the code-graphs that represent each method. These code-graphs are derived from the parse tree formed by that method. See Aris details.

I was senior scientific officer for the Dr Inventorproject, discovering creative analogical comparisons between research publications in computer graphics. See the creative analogies between SIGGRAPH graphics publications, also analogies between ICCC creativity papers are available using a simpler interface. See Maynooth's Dr Inventor web-page.

Evolutionary Optimisation with Extra-Mendelian Inheritance. We used ancestral caches of genomic data for evolutionary algorithms and optimisation. Our (analogical) inspiration came from a controversial paper in Nature by Lolle et al (2005) and top of the "All Time Top 10" rankings (jan 2020) of the prestigious Faculty of 1000.


Research Group:
Postgrad levelMengqi Hao, Sarbojit Bhattacharjee, Saad Khalid, Wingyin Ha, Xiaoxu Tan. Undergraduate research students: (23-24) Daman Mannering, Levis Thomas, Danielus Grebliauskas, Olatomiwa.Omotoso, Rohit Dina Nath.
My students look at many topics related to discovering and exploiting hidden and latent similarity between lexically based artifacts, such as documents and English and source code in various languages including Java, C# and Python. Specifically; knowledge graphs and Open Information Extraction, parser driven Open IE (2) knowledge graph processing and elaboration (3) graph matching algorithms like VF3, ISMAGS (4) Information extraction from source code (C#, Python, Java, ...) (5) Retrieval Augmented Generation (RAG). This work follows on from the Cre8blend and Aris projects above.

Diarmuid's Publications

Some Program Committee Membership: 15th International Conference on Computational Creativity ICCC Sweden, June 2024. IJCAI 2024 AI, Arts & Creativity special track.

My GitHub repo, but it only contains major updates.

Teaching

Undergraduate courses I currently teach (bold) or have taught:

CS404- Artificial Intelligence and Natural Language Processing, CS430 - Advanced Concepts: Computational Creativity, CS101- Introduction to Programming, CS130- Databases, CS142- Introduction to Computer Science, CS401- Machine Learning and Neural Networks, CS404 - Computational Linguistics, CS355- Artificial Intelligence, CS335 Software Engineering, Expert Systems, CS120 - End User computing, CS102 - Introduction to Computer Systems, Digital Logic Design.

MSc in Computer Science (Software Engineering) - :

Edu-tainment to underpin the fundamentals of variable manipulation. Order from Chaos is online or download the standalone version.

CS607 Requirements Engineering and Systems Design using UML.




View Maynooth in a larger map

Search within nuim.ie, or add my NUIM Google Search Gadget to your page.

URL: http://www.cs.nuim.ie/~dod/
Revised: 2022.
Contact: webmaster@cs.nuim.ie
Copyright 2016 Maynooth University, Ireland.