<?xml version="1.0" encoding="UTF-8"?><xml><records><record><source-app name="Biblio" version="7.x">Drupal-Biblio</source-app><ref-type>47</ref-type><contributors><authors><author><style face="normal" font="default" size="100%">Smeaton, Alan F.</style></author><author><style face="normal" font="default" size="100%">Graham, Yvette</style></author><author><style face="normal" font="default" size="100%">McGuinness, Kevin</style></author><author><style face="normal" font="default" size="100%">O'Connor, Noel E.</style></author><author><style face="normal" font="default" size="100%">Quinn, Seán</style></author><author><style face="normal" font="default" size="100%">Arazo Sanchez, Eric</style></author></authors></contributors><titles><title><style face="normal" font="default" size="100%">Exploring the Impact of Training Data Bias on Automatic Generation of Video Captions</style></title><secondary-title><style face="normal" font="default" size="100%">MultiMedia Modeling</style></secondary-title></titles><dates><year><style  face="normal" font="default" size="100%">2019</style></year></dates><publisher><style face="normal" font="default" size="100%">Springer International Publishing</style></publisher><pub-location><style face="normal" font="default" size="100%">Cham</style></pub-location><pages><style face="normal" font="default" size="100%">178–190</style></pages><isbn><style face="normal" font="default" size="100%">978-3-030-05710-7</style></isbn><language><style face="normal" font="default" size="100%">eng</style></language><abstract><style face="normal" font="default" size="100%">&lt;p&gt;A major issue in machine learning is availability of training data. While this historically referred to the availability of a sufficient volume of training data, recently this has shifted to the availability of sufficient unbiased training data. In this paper we focus on the effect of training data bias on an emerging multimedia application, the automatic captioning of short video clips. We use subsets of the same training data to generate different models for video captioning using the same machine learning technique and we evaluate the performances of different training data subsets using a well-known video caption benchmark, TRECVid. We train using the MSR-VTT video-caption pairs and we prune this to reduce and make the set of captions describing a video more homogeneously similar, or more diverse, or we prune randomly. We then assess the effectiveness of caption-generating trained with these variations using automatic metrics as well as direct assessment by human assessors. Our findings are preliminary and show that randomly pruning captions from the training data yields the worst performance and that pruning to make the data more homogeneous, or diverse, does improve performance slightly when compared to random. Our work points to the need for more training data, both more video clips but, more importantly, more captions for those videos.&lt;/p&gt;
</style></abstract></record><record><source-app name="Biblio" version="7.x">Drupal-Biblio</source-app><ref-type>47</ref-type><contributors><authors><author><style face="normal" font="default" size="100%">Quinn, Seán</style></author><author><style face="normal" font="default" size="100%">Mileo, Alessandra</style></author></authors></contributors><titles><title><style face="normal" font="default" size="100%">Towards Architecture-Agnostic Neural Transfer: a Knowledge-Enhanced Approach</style></title><secondary-title><style face="normal" font="default" size="100%">Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence, {IJCAI-19}</style></secondary-title></titles><dates><year><style  face="normal" font="default" size="100%">2019</style></year><pub-dates><date><style  face="normal" font="default" size="100%">7</style></date></pub-dates></dates><urls><web-urls><url><style face="normal" font="default" size="100%">https://doi.org/10.24963/ijcai.2019/915</style></url></web-urls></urls><publisher><style face="normal" font="default" size="100%">International Joint Conferences on Artificial Intelligence Organization</style></publisher><pages><style face="normal" font="default" size="100%">6452–6453</style></pages><language><style face="normal" font="default" size="100%">eng</style></language></record><record><source-app name="Biblio" version="7.x">Drupal-Biblio</source-app><ref-type>47</ref-type><contributors><authors><author><style face="normal" font="default" size="100%">Quinn, Seán</style></author><author><style face="normal" font="default" size="100%">Murphy, Noel</style></author><author><style face="normal" font="default" size="100%">Smeaton, Alan F.</style></author></authors></contributors><titles><title><style face="normal" font="default" size="100%">Tracking Human Behavioural Consistency by Analysing Periodicity of Household Water Consumption</style></title><secondary-title><style face="normal" font="default" size="100%">2nd International Conference on Sensors, Signal and Image Processing (SSIP 19)</style></secondary-title></titles><keywords><keyword><style  face="normal" font="default" size="100%">Ambient Assisted Living</style></keyword><keyword><style  face="normal" font="default" size="100%">Home Monitoring</style></keyword><keyword><style  face="normal" font="default" size="100%">Internet of Things</style></keyword><keyword><style  face="normal" font="default" size="100%">Sensor Applications</style></keyword><keyword><style  face="normal" font="default" size="100%">Sensor Networks</style></keyword></keywords><dates><year><style  face="normal" font="default" size="100%">2019</style></year></dates><publisher><style face="normal" font="default" size="100%">ACM</style></publisher><pub-location><style face="normal" font="default" size="100%">Prague, Czech Republic</style></pub-location><language><style face="normal" font="default" size="100%">eng</style></language><abstract><style face="normal" font="default" size="100%">&lt;p&gt;People are living longer than ever due to advances in healthcare, and this has prompted many healthcare providers to look towards remote patient care as a means to meet the needs of the future. It is now a priority to enable people to reside in their own homes rather than in overburdened facilities whenever possible. The increasing maturity of IoT technologies and the falling costs of connected sensors has made the deployment of remote healthcare at scale an increasingly attractive prospect. In this work we demonstrate that we can measure the consistency and regularity of the behaviour of a household using sensor readings generated from interaction with the home environment. We show that we can track changes in this behaviour regularity longitudinally and detect changes that may be related to significant life events or trends that may be medically significant. We achieve this using periodicity analysis on water usage readings sampled from the main household water meter every 15 minutes for over 8 months. We utilise an IoT Application Enablement Platform in conjunction with low cost LoRa-enabled sensors and a Low Power Wide Area Network in order to validate a data collection methodology that could be deployed at large scale in future. We envision the statistical methods described here being applied to data streams from the homes of elderly and at-risk groups, both as a means of&amp;nbsp; early illness&amp;nbsp; detection&amp;nbsp; and&amp;nbsp; for&amp;nbsp; monitoring&amp;nbsp; the well-being of those with known illnesses.&lt;/p&gt;
</style></abstract></record></records></xml>