Threat intelligence, artificial life, AI to predict medical capacity demand and a 3D printed mechatronic prosthetic were just some of themes and projects in the 2019 Dublin City University final year showcase for the Faculty of Engineering and Computing.
“Today, we have the opportunity to recognise and celebrate the work of these students. We all go through many stages in our careers and lives, and today, for these students, marks an important milestone. Today, they display this major highlight of their academic achievements for their peers, their academic mentors, and for the many guests who join us from industry and the wider DCU community,” said Professor Lisa Looney, Dean, Faculty of Engineering and Computing, DCU.
“The range of final year projects reflects the great opportunities for these young graduates to contribute to the advancement of Irish society over the coming years. Many of the projects displayed represent products with immediate commercial potential; all demonstrate ingenuity, skill and technical capability of the highest calibre.”
With more than 230 projects on display, covering a diverse range of topics and showcasing a depth of creativity and ingenuity that bodes well for future prospects, the 2019 showcase was supported by the Openet, Fidelity, FINEOS, Davy Group and IBM, with a range of alumni and prospective employers also in attendance.
David Jenkins leveraged a threat intelligence database to develop a tool to perform a hybrid analysis, providing visualisations for distribution of threats, incidence, and various other data aspects.
Eileen McGarvey’s “Aithin Canúint” used the Hidden Markov Model toolkit (HTK) to create a system to discern between major Irish dialects, such as between Donegal, Galway and Cork accents in the Irish language. With potential applications in call centre and customer service areas, McGarvey also sees potential in educational applications.
Chloe D’Arcy and Amy Rowe used AI used to determine the length of stay of hospital patients, with implications for staffing, resourcing and capacity planning.
Key data points are assessed to predict the likely stay of patients entering the system. Algorithms have been developed and implementation will be based on Python Django. It can provide visualisations such as heat maps for admissions. While the pair admit this is merely a proof of concept and not a working system, the potential is clear. Unsurprisingly, both of the graduates have already secured jobs and so may not be able to persevere with the system to fruition.
Khaled Alaydi challenged himself to create a prototype mechatronic prosthetic arm for less than the commercial €20,000 options available. His prototype weighs 600g and was produced around a Teensy Board controller with muscle impulse sensors, for around €200.
Marin Bivol’s Java Script operating system is a performance revelation and aimed at developers to allow them to develop an application once and have it run anywhere. His working example enjoys impressive performance without GPU acceleration, though can benefit from that boost too.
A Fake News Detector from David Talan uses various techniques, such as comparative analysis, to score how likely a content piece is to be fake news. His working system scans other sources for incidence of the story, checks for certain words, and uses a passive aggressive ML classifier. A weighting system for various data points combines for an overall score.
Talan admitted that accuracy may be affected with a genuine scoop, but the reputation of the source prior to the scoop would be a key indicator in such circumstances.
David Craig’s “Artificial Life” uses a genetic algorithm to perform environment modelling. Written entirely from scratch, Craig said future versions will have more detailed levels of physical characteristics to increase the complexity of the modelling.
A key shortcoming of renewable power is fluctuation in the sources. Odhran Casey assessed the viability of storage systems, in the context of the home, for renewable power.
TechCentral Reporters
Ltd
