Irish Conference on Artificial Intelligence and Cognitive Science

December 6-7th 2018

Trinity College Dublin


The conference will be held in the Trinity Long room Hub on the main campus, Trinity College Dublin. The detailed program will be published after the paper review process for this year's event. A draft outline program is presented here to facilitate travel planning by attendees.

Schedule - Thursday 6th December (Full Day)

9:30 - 9:45Welcome
09:45 - 10:30 Keynote by Prof Dr Torsten Schaub, University of Potsdam, Germany

Answer Set Programming in a Nutshell
Answer Set Programming (ASP) has become a popular approach to declarative problem solving. More precisely, ASP is a rule-based formalism for modeling and solving knowledge-intense combinatorial (optimization) problems. What makes ASP attractive is its combination of a declarative modeling language with highly effective solving engines. This allows us to concentrate on specifying - rather than programming the algorithm for solving - a problem at hand. Historically, ASP has its roots in deductive databases, logic programming, and non-monotonic reasoning; its solving engines draw on the same technology as solvers for satisfiability testing. Given this origin, ASP is tailored to support closed as well as open world reasoning, which makes it predestined for knowledge representation and reasoning tasks. Interesting applications of ASP can be found in decision support systems, industrial team-building, music composition, natural language processing, product and software configuration, phylogeneticics, robotics, systems biology, timetabling, and many more. The talk will give a gentle introduction to ASP, its logical foundations, modeling capabilities, and solving engines, and conclude with an outlook on the ASP's potential impact as a knowledge-driven AI tool.

10:30-11:00Coffee Break
11:00-12:30Session 1 - Numerical Reasoning and Algorithms
  • Logical Reasoning and Mathematics Skills in Children and Adults
  • Kinga Morsanyi, Teresa McCormack and Eileen O'Mahony
  • Kernel Methods for Time Series Classification and Regression
  • Mourtadha Badiane, Martin O'Reilly and Padraig Cunningham
  • The role of horizontal, vertical and sagittal axes on visuo-spatial number mapping in children
  • Sarah M. Cooney, Corinne A. Holmes and Fiona N. Newell
  • A qualitative investigation of the explainability of defeasible argumentation and non-monotonic fuzzy reasoning
  • Lucas Rizzo and Luca Longo
  • Meta-Learning for Recommendation Algorithm Selection
  • Andrew Collins, Joeran Beel and Dominika Tkaczyk
  • A Survey of k-NN Methods for Time Series Classification and Regression
  • Vivek Mahato, Martin O'Reilly, and Padraig Cunningham
12:30-13:30Lunch Break and Poster Session
13:30-15:00Session 2 - AI and Reasoning Applications
  • A Comparison of Agent-Based Models and Equation Based Models for Infectious Disease Epidemiology
  • Elizabeth Hunter, Brian Mac Namee and John Kelleher
  • Vehicular traffic flow intensity detection and prediction through mobile data usage
  • Maurice Saliba, Charlie Abela, and Colin Layeld
  • Counterfactual and semi-factual thoughts in moral judgements about failed attempts to harm
  • Mary Parkinson and Ruth M. J. Byrne
  • Early fusion of Dense Optical Flow with Image for Semantic Segmentation in Autonomous Driving
  • Prashanth Viswanath, Ganesh Sistu, Mihai Ilie, Senthil Yogamani and Jonathan Horgan
  • She just doesn't understand me! Curing Alexa of her Alexithymia
  • Eoghan Furey and Juanita Blue
  • Manipulating Moral Dumbfounding: Inhibiting the Identification of Reasons
  • Cillian McHugh, Marek McGann, Eric R. Igou, and Elaine L. Kinsella
15:00-15:30Coffee Break and Posters
15:30-17:00Session 3 - Big Data and Machine Learning Applications
  • MeetupNet Dublin: Discovering Communities in Dublin's Meetup Network
  • Arjun Pakrashi, Elham Alghamdi, Brian Mac Namee and Derek Greene
  • Day Ahead Forecasting of FAANG Stocks Using ARIMA, LSTM Networks and Wavelet Decomposition.
  • Tom Skehin, Martin Crane, Marija Bezbradica
  • Machine learning EEG to predict Cognitive Functioning in Multiple Sclerosis Patients and Controls
  • H. Kiiski, L. Jollans, S. O Donnchadha, H. Nolan, R. Lonergan, S. Kelly, M.C. O'Brien, K. Kinsella, J. Bramham, T. Burke, M. Hutchinson, N. Tubridy, R.B. Reilly, R. Whelan
  • ParsRec: Recommending Bibliographic Reference Parsing Algorithms Using Meta-Learning
  • Dominika Tkaczyk, Paraic Sheridan and Joeran Beel
  • An Empirical Comparison of Syllabuses for Curriculum Learning
  • Mark Collier and Joeran Beel
  • Evaluating Load Adjusted Learning Strategies for Client Service Levels Prediction from Cloud-hosted Video Servers
  • Obinna Izima, Ruairi de Frein and Mark Davis
17:00Wrap Up

Conference Dinner O'Callaghan Davenport Hotel, Dublin

Google Maps

Schedule - Friday 7th December (Half Day)

09:15-10:15Session 4 - Understanding Data
  • Heterogeneous Multi-Agent Deep Reinforcement Learning for Traffic Lights Control
  • Jeancarlo Arguello Calvo and Ivana Dusparic
  • Milan: Automatic Generation of R2RML Mappings
  • Sahil Mathur, Declan O'Sullivan and Rob Brennan
  • Mr. DLib's Architecture for Scholarly Recommendations-as-a-Service
  • Joeran Beel, Andrew Collins and Akiko Aizawa
  • Generating Diverse and Meaningful Captions
  • Annika Lindh, Robert J. Ross, Abhijit Mahalunkar, Giancarlo Salton and John D. Kelleher
10:15 - 11:00 Keynote by Prof Dr Monica Bucciarelli, University of Turin, Italy
Informal Algorithms in Children
Where does the ability to devise algorithms come from, and on what does it depend? Most cognitive abilities appear to depend on interactions between innate factors and experience. My studies of children, in collaboration with colleagues (Robert Mackiewicz, Sangeet Khemlani, and Phil Johnson-Laird) show that fifth-grade children (10-11 year olds) who have had no experience with programming are nevertheless able to understand informal algorithms and even to devise their own. The domain concerned re-arranging the order of toy cars in a railway train, using a single track and a siding. Such a track allows for powerful algorithms, since both the siding and one side of the track act as places to “store” cars during re-arrangements. These studies led us to three main conclusions. First, children can devise algorithms, even those that are recursive (i.e., they depend on repeated loops of operations), although they are harder than those that are not recursive (i.e., they depend on a short list of operations). The former place a greater load on working memory than the latter. Second, children differ in ability, though there is no reliable difference between boys and girls. Ability is likely to reflect a difference in the processing capacity of working memory (at least partially determined by innate factors). So, for example, when children cannot even touch the cars, their “iconic” gestures of moves help them. When they are prevented from gesturing, their performance is poorer. Third, children (and adults) use kinematic mental simulations to envisage the effects of moves on the railway track.
11:00-11:30Coffee Break & Poster Session
11:30-13:15Session 5 - Speech, Video and Reasoning
  • Electrophysiological Correlates of Semantic Dissimilarity Reflect the Comprehension of Natural, Narrative Speech
  • Michael P. Broderick, Andrew J. Anderson, Giovanni M. Di Liberto, Michael J. Crosse, Edmund C. Lalor
  • Comparing Joint Speech and Synchronous Speech: What Happens When We Add More Speakers?
  • Fred Cummins
  • Exploring a Perceptually-weighted DNN-based Fusion Model for Speech Separation
  • Alessandro Ragano, Andrew Hines
  • Are naive and scientific theories accessed by reasoning processes sequentially or in parallel?
  • Aidan Feeney and Eoin Travers
  • The sound of silence: how traditional and deep learning based voice activity detection influences speech quality monitoring
  • Rahul Jaiswal, Andrew Hines
  • ADNet: A Deep Network for Detecting Adverts
  • Murhaf Hossari, Soumyabrata Dev, Matthew Nicholson, Killian McCabe, Atul Nautiyal, Clare Conran, Jian Tang, Wei Xu and Francois Pitie
  • An Interpretable Machine Vision Approach to Human Activity Recognition using Photoplethysmograph Sensor Data
  • Eoin Brophy, Jose Juan Dominguez Veiga, Zhengwei Wang, Alan F. Smeaton and Tomas Ward
13:15-13.30Closing Remarks
13.30Conference closed for attendees
13:30-14:30AIAI Community Meeting (invitation-only boardroom session)