Innovator

William Ikenna-Nwosu  • Peerlist

<p>Collaborated in a team of three on a project for our Distributed Systems (COMP30220) module where we created Dexter; a code reviewer web application where players are shown code snippets from other players, and they can find possible errors in the code and make suggestions on how to improve it.</p><h3>Deliverables</h3><ul><li><p><strong>Wrote comprehensive final report (10 pages)</strong> in collaboration with team members detailing the technologies used and system created to architect distribution under the following headings complete with diagrams and explanations; What is the application domain?, What will the application do?, What are the key components and how will they interact?, Technology stack, System Overview, Reflections (What were the key challenges you faced in completing the project?, How did you overcome them?, What would you have done differently if you could start again?, What have you learnt about the technologies you have used?, Limitation, Benefits)</p></li></ul>

Dexter

<p>Collaborated in a team of four on a project for our Software Engineering III (COMP30050) module where we created Scentaur, a Java code smell detector that analyses source code and gives a statistical report of the code smells detected across the codebase. </p><h3>Responsibilities</h3><p>I was assigned the role of Lead Designer and was responsible for the following:</p><ul><li><p><strong>Leading discussion surrounding conceptual design </strong>and creating overall architecture design for the code smell detector by breaking concepts down into classes using OOP principles and drafting hierarchies for the system as a whole.</p></li><li><p><strong>Determining relationships between classes and designing interfaces</strong> with teammates to make software more flexible when features need to be extended and to make software integration easier in the long run.</p></li><li><p><strong>Explaining possible methods to implement concepts</strong> to teammates and getting their feedback on them iteratively to eventually reach a conclusion.</p></li><li><p><strong>Estimating deadlines</strong> in which we could implement parts of the project and organising our collaboration strategy around that.</p></li><li><p><strong>Discussing design decisions </strong>with teammates to make an optimal choice as a collective based on the length of time to implement and the features and parameters we prioritised most for our software product.</p></li></ul><h3>Deliverables</h3><ul><li><p><strong>Wrote comprehensive interim report (15 pages)</strong> in collaboration with team members detailing project plan and delegation of labour including; Introduction, Specification, Software Overview, Major Responsibilities and Work Breakdown, Team Communication, Concluding Remarks, Acknowledgements and References.</p></li><li><p><strong>Wrote comprehensive final report (20 pages) </strong>in collaboration with team members detailing specifics of project outcome including; Introduction, System Features, Project Design, Successes and Failures, Team Communication, Project In Detail, Acknowledgements and References. Made use of Google Docs to aid collaboration and MS Office Suite (Word, Excel, PowerPoint and Office) during writing of interim and final reports.</p></li></ul>

Scentaur

<p>Completed my thesis under the supervision of Colm O'Riordan over the course of one module Artificial Intelligence Project (CT5129) at 30 ECTS.</p><p>My self-proposed thesis is titled <em>"GAPRS: Graph-based Academic Paper Recommendation System"</em> and focussed on creating a graph and network based academic paper recommender system via the OpenAlex API and made use of generative AI to incorporate HCI, XAI and aid interpretation of data visualizations generated by GAPRS which come in the form of graphs and networks. </p><h3>Deliverables</h3><ul><li><p>Submitted 68-page thesis to examiners for review.</p></li></ul><h3>Outputs</h3><ul><li><p>CLI tool which integrates the OpenAlex API for academic papers and ChatGPT for XAI and HCI.</p></li><li><p>Market research and competitor analysis of research paper recommender systems (RPRS).</p></li><li><p>Historical survey of research paper recommender systems.</p></li><li><p>Literature review consisting of the following: identifying consensus in the RPRS space, identifying and summarizing problems facing RPRS community, identifying and collecting open and emerging challenges in the RPRS community, identifying and collecting open challenges in the graph learning recommender systems community, identifying relevant work in related research fields.</p></li><li><p>Compilation of the following: List of known recommender system evaluation metrics to date, list of relevant network science metrics, list of known recommendation algorithms to date, list of direct and indirect competitors to GAPRS.</p></li><li><p>Novel methodology for approaching problem of recommending academic papers to users via creation of new network science metric, avg(NCCC, BCCC). As well as creation of hybrid weighted undirected 1-Degree and 1.5-Degree multi-egocentric citation networks.</p></li></ul><h3>Outcomes</h3><ul><li><p>Awarded Microsoft University Mentorship 2023 which transformed my thesis into an industry-research collaboration.</p></li></ul>

Masters Thesis: GAPRS

<p>Completed my thesis under the supervision of Félix Balado Pumariño and moderation of Shen Wang and Damian Dalton, over the course of two modules totalling 15 ECTS due to University College Dublin (UCD)’s new Final Year Project (FYP) structure: FYP Foundations (COMP30900) at 5 ECTS and FYP Design and Implementation (COMP30910) at 10 ECTS. 

My thesis is titled "Finding the Minimum Number of Sudoku Clues Through Information Theory" and focussed on deriving an information theoretic explanation for the results at the state-of-the-art (SOTA) of the Minimum Sudoku Clue Problem (MSCP). That explanation would then be used to lay the foundation for improvements beyond the SOTA of the MSCP.</p><h3>Deliverables</h3><ul><li><p>FYP Foundations (COMP30090): Submitted two deliverables as part of requirements consisting of a 28-page foundations report and 10-minute presentation which both summarized my plan of action and were presented to both moderators. </p></li><li><p>FYP Design and Implementation (COMP30910): Submitted 39-page thesis to examiners for review. </p></li></ul><h3>Outputs</h3><ul><li><p>Historical survey of Sudoku and MSCP from origins to SOTA.</p></li><li><p>Analytical proof of capacity of N-ary erasure communication channels for Sudoku codes.</p></li><li><p>Laid groundwork for potential advancements in error-free communication across the channel.</p></li><li><p>Information-Theoretic rates of 4x4, 9x9, 16x16 and 25x25 Sudoku grids.</p></li><li><p>Identified avenues of future research to build on thesis.</p></li></ul><h3>Outcomes</h3><ul><li><p>I was awarded the Franz Geiselbrechtinger Medal 2019/2020 for having the best thesis in my graduating cohort.</p></li></ul>