Welcome to the School of Computer Science and Information Technologies Seminar Series

"We can only see a short distance ahead,
but we can see plenty there that needs to be done."
~ Alan Turing, Computing machinery and intelligence

Our Next Speaker

The Seminar Series is now finished for 2019.
The Series will start back on the 8th of Jan 2020.

About CSIT Seminar Series

Upcoming Seminars

Past Seminars

Register Your Interest

About

The CSIT seminar series aims at creating opportunities to connect with colleagues and to focus on shared interests in both the research and practice. We hope to foster collaboration, the creation of new ideas and to achieve better visibility for members of the CSIT research community in UCC.

Our all-inclusive approach allows for a broad range of topics and backgrounds and we hope to bring to light the great many ways the CS research community is impacting society through research and teaching.

The seminar series is a great opportunity for any CS staff and CS postgraduate students to engage with colleagues and share their latest findings, practice a conference talk, review their domain or give a tutorial, among other things. Attendance is welcome from the School staff members and students.

If you are interested in participating in the seminar series, please fill out our registration form

If you ave any feedback (good or bad), we would love to hear from you. You can use our feedback form.

This series is BYOB (Bring Your Own Beverage), we will try to bring biscuits.

If you have questions, feel free to contact us. You can find our contact details by clicking here.

Upcoming Seminars

Date, Time, Place, Talk Type Speaker, Position, Affiliation Title, Abstract, Keywords
2020
08/Jan/2020 13:00-14:00, TBC, Cathal Hoare, Senior Postdoc To be Confirmed
15/Jan/2020 13:00-14:00, TBC, Conference paper rehearsal Steven Prestwich, Lecturer, UCC Title: Bayesian Constraint Acquisition
Abstract: Modeling a combinatorial problem is a hard and error-prone task requiring expertise. Constraint acquisition methods can automate this process by learning constraints from examples of solutions, and usually non-solutions. I describe a new statistical approach based on sequential analysis (sequential Bayesian hypothesis testing) that is orders of magnitude faster than existing methods. It is also robust in the sense that it can learn constraints correctly even when the data contain many errors.
Keywords: AI, constraint acquisition, Bayesian
22/Jan/2020 13:00-14:00, TBC, Erica Volta, Postdoc, University of Genova To be Confirmed
29/Jan/2020 13:00-14:00, TBC, Ahmed Zahran, Lecturer, UCC To be Confirmed
05/Feb/2020 13:00-14:00, TBC, Available
12/Feb/2020 13:00-14:00, TBC, Available
19/Feb/2020 13:00-14:00, TBC, Available
26/Feb/2020 13:00-14:00, TBC, Available
04/March/2020 13:00-14:00, TBC, Available
11/March/2020 13:00-14:00, TBC, Available
18/March/2020 13:00-14:00, TBC, Available
25/March/2020 13:00-14:00, TBC, Available
01/Apr/2020 13:00-14:00, TBC, Available

Past Seminars

Date, Time, Place, Talk Type Speaker, Position, Affiliation Title, Abstract, Keywords
12/Dec/2019 13:00-14:00, WGB - 402, Presentation of an initiative on smart cities
Vanessa Pulgarin, Research Support Officer, Cork Smart Gateway @UCC
Title: Cork Smart Gateway - Innovating, creating, connecting
Abstract: The Cork Smart Gateway initiative is composed by members from Cork City Council, Cork County Council, UCC, CIT and Tyndall. It aims to enhance the reputation of Cork as an attractive place to live, work, visit and invest. A place where public infrastructure and public service provision utilise best in class technology solutions and one which allows all stakeholders to participate in decision making and enjoy an enhanced environment. This initiative will compliment and support the already large number of organisations in Cork who are working on smart solutions to improve the physical realm, from energy conservation to open data.
Keywords: Smart cities, test beds, IoT
Click here to view presentation
06/Dec/2019 13:00-14:00, WGB - 402, Review of a domain & Work in progress
Anne-Marie George, Postdoc, TU Berlin, Efficient Algorithms Group
Title: Liquid Democracy with Ranked Delegations
Abstract: I will first give an introduction to Computational Social Choice and Voting Theory to set the context for discussing Liquid Democracy. Liquid Democracy is a voting framework in which voters can decide to either vote on an issue directly or to delegate their vote to other voters. As such it can be seen as a hybrid between Direct Democracy and Representative Democracy. While Liquid Democracy aims at overcoming certain problems of these two more traditional types of democracy, it brings it's own types of problems with it. As theoretical computer scientists we are interested in formalising desiderata (e.g., that voters cannot manipulate the system) and design mechanisms that enforce these. An analysis of these algorithms for Liquid Democracy typically comes with a graph theoretical flavour. I will present work in progress on the specific case, where voters are allowed to specify several delegation possibilities in form of a ranking.
Keywords: Computational Social Choice, Voting Theory, Liquid Democracy, Vote Delegation, Ordinal Preferences
Click here to view presentation
27/Nov/2019 14:00-15:00, Electronic Engineering building room L3/EDA Lab, Industry Rajeev Jain, Emeritus Professor and Senior Director, Qualcomm Title: Qualcomm AI Research
Abstract: The talk will review main directions in corporate Qualcomm R&D in AI with examples of successful products and genuine technical innovations.
21/Nov/2019 13:00-14:00, WGB - 402, Technical piece of work Dimitrios Zormpas, Marie Curie EDGE Fellow, Tyndall National Institute Title: From ALOHA-based to SYNC-based LoRa communications
Abstract: Due to the ALOHA-style transmission policy of LoRa networks, delivering a large number of packets by a large number of nodes becomes impossible. To avoid bursts of collisions and expedite data collection, we propose scheduling of transmissions in time slots. We provide evidence through extensive experiments that time-slotted synchronous LoRa communications are feasible under different node arrangements.
Keywords: wireless communications, LPWAN, LoRa, Scheduling
07/Nov/2019 13:00-14:00, WGB - 402, Technical piece of work Marc van Dongen, Lecturer, UCC Title: Implementing a CSP Solver
Abstract: This talk presents work about a general-purpose, sound and complete solver for deciding and enumerating the solutions of binary Constraint Satisfaction Problem (CSP) Instances which are provided in extension.
Keywords: Binary CSPs, Solver
31/Oct/2019 13:00-14:00, WGB - 402, Event presentation & recruitment
Bastien Pietropaoli & Milan De Cauwer, Post-docs, Insight@UCC Confirm@UCC
Title: IrlCPC / AIPO - Post-mortem & How we created the problem set
Abstract: In this talk, we will present the results of IrlCPC (Irish Collegiate Programming Contest) and AIPO (All-Ireland Programming Olympiads) that were held at UCC on March 19 2019. We will present how the candidates/teams performed. We will explain how we created a fun but interesting problem set and how our various test cases reflected in the scoreboards. If you would be interested in participating in the writing of the problem set for 2020, we would strongly encourage you to come to this talk.
Keywords: Programming Competition IrlCPC - AIPO
Click here to view presentation
25/Oct/2019 13:00 - 14:00, L2 Electrical and Electronic Engineering Building UCC Dr. Michele Magno, Associate professor, ETH Zurich Title: Smart Sensing in the IoT Era: Machine Learning on Ultra Low Power Microcontrollers
Abstract: Machine Learning (ML) and artificial intelligence are pervading the digital society. Today, even low power embedded systems are incorporating ML, becoming increasingly "smart." This talk gives an overview of ML methods and algorithms to process and extract useful near-sensor information in end-nodes of the "internet-of-things", using low-power microcontrollers ARMCortex-M; Bluetooth low energy SoC, and other low power microcontrollers. The main objective is to show how Machine Learning algorithms can be adapted to the performance constraints and limited resources of low-power microcontrollers both the popular ARM-Cortex-M family and novel Parallel Ultra Low Power (PULP) family based on RISC-V cores. Finally, the talk will introduce an open-source library based on a Fast Artificial neural network, that allows implementing quickly Artificial Neural Network on ARM Cortex-M Family and PULP. The talk will also show with use-case examples, the benefits of energy-efficient ML to achieve self-sustaining smart sensing.
Keywords: AI, Machine Learning, Internet of Things, energy efficiency
17/Oct/2019 13:00-14:00, WGB - 402, Overview of Research Projects
Dirk Pesch, Professor, ADVANCE CRT@UCC
Title: Does the Internet of Things exist only in silos? Challenges for interoperability
Abstract: The Internet of Things, that is connecting the physical world to the Internet to enable smart objects and an unprecedented access to data about our world, has been around for over 20 years, since Kevin Ashton of MIT coined the phrase in 1999. However, much of its promise, both technical and from a commercial perspective, is still to be fulfilled. Recently estimates of the number of IoT devices has been downgraded from 50 billion by 2020 (prediction in 2013) to around 6 - 7 billion by 2022. Why is it that we are not seeing the promise of the Internet of Things envisioned a few years ago? This talk will look one of the issues underlying the much slower than expected growth, the issue of interoperability. Firstly, we will summarise what the internet of things is, then present a brief overview of the different technology components, which will highlight the issues. We then suggest a number of mechanisms that may help to overcome the siloed situation of the current Internet of Things. These include interoperability platforms, software defined infrastructure, interoperable data semantics and the potential of artificial intelligence.
Keywords: IoT, Future of IoT, Interoperability, Software defined infrastructure, Artificial Intelligence
03/Oct/2019 13:00-14:00, WGB - 402, Tutorial / Research Strategies
Klaas-Jan Stol, Lecturer, Lero@UCC
Title: The ABC of Software Engineering Research
Abstract: A variety of research methods and techniques are available to SE researchers, and while several overviews exist, there is consistency neither in the research methods covered nor in the terminology used. Furthermore, research is sometimes critically reviewed for characteristics inherent to the methods. We adopt a taxonomy from the social sciences, termed here the ABC framework for SE research, which offers a holistic view of eight archetypal research strategies. ABC refers to the research goal that strives for generalizability over Actors (A) and precise measurement of their Behavior (B), in a realistic Context (C). The ABC framework uses two dimensions widely considered to be key in research design: the level of obtrusiveness of the research and the generalizability of research findings. We discuss metaphors for each strategy and their inherent limitations and potential strengths. We illustrate these research strategies in two key SE domains, global software engineering and requirements engineering, and apply the framework on a sample of 75 articles. Finally, we discuss six ways in which the framework can advance SE research.
Keywords: Research Strategies Empirical Research
Click here to view presentation
19/Sept/2019 13:00-14:00, WGB - 402, Workshop rehearsal
Guillaume Escamocher, Post-doc, Insight@UCC
Title: Solving Logic Grid Puzzles with an Algorithm that Imitates Human Behavior
Abstract: I will present in this seminar our entry to the Workshop on Progress Towards the Holy Grail Challenge. The Challenge consists in providing an automated process that can solve the puzzles from the logicgridpuzzles.com website, as well as explain how it got to the solution. The approach used by our algorithm mimics the way a human would try to solve the same problem. Every progress made during the solving process is accompanied by a detailed explanation of our program’s reasoning. Since this reasoning is based on the same heuristics that a human would employ, the user can easily follow the given explanation.
Keywords: Artificial Intelligence Holy Grail of Programming Explanation Logic Grid Puzzle
Click here to view presentation
12/Sept/2019 13:00-14:00, WGB - 402, Technical piece of work Laura Maye, Lecturer, School of CS & IT Title: Grassroots Radio: Radio for the 21st Century
Abstract: How should radio work for remote and rural communities in light of urbanisation and globalisation, which are diminishing local voices? Grassroots Radio aims to enable community members, through the use of radio as a grassroots peer production platform, to have a voice on local issues, topics, and concerns of value to them. The project is employing an open platform called RootIO: founded in 2013, RootIO can be used by community members to broadcast over FM using any basic phone. Currently, RootIO is deployed in six rural and remote communities across Ireland, including Bere Island and the Islands on the southwest of Ireland, two communities in the Island of Madeira in Portugal, and two communities in the Danube Delta in Romania. In this talk, I will present work completed in the project to date, with the goal of highlighting the RootIO intends to enable communities low-barrier means of communicating at a hyper local level. This will also be accompanied by a demonstration of the platform.
Keywords: Peer-to-peer production Communication Community Rural
06/Sept/2019 12:00-13:00, WGB - 2.26, Overview of research projects Lara Quijano Sánchez, Associate Lecturer, Autónoma University of Madrid Title: The Associate: From Social Group Recommenders to Active Citizen Participation
Abstract: This talk is an account Lara's research career, presenting past, present and future research projects. It contains an overview of several research projects related to i) Recommender systems and the hypothesis that social factors and organizations such as personality and tie strength have a positive impact in group recommendation processes; ii) Social network analysis and decision support systems for client capture applied to social media and financial networks; iii) Applications of data science in policing including VeriPol, an Investigation support tool in the Spanish National Police that allows to identify false reports, and HaterNet, an intelligent system in the Spanish National Office Against Hate Crimes that monitors and visualizes hate speech in Social Media. iv) VioGen: a study of a copycat effect in intimate partner homicides and definition of a risk model, a lesson from Spanish femicides. v) Active citizen participation: explanatory recommendation systems for the improvement and motivation of citizen participation.
Keywords: AI Social Recommender Systems Tie Strength Social Network Analysis Sentiment Analysis Natural Language Processing Predictive Policing E-participation Explainable AI
28/Aug/2019 12:00-13:00, WGB - G09, Conference rehearsal
Brian McCarthy, Ph.D Student, Connect@UCC
Title: OpenCV2X Mode 4: A Simulation Extension for Cellular Vehicular Communication Networks
Abstract: In this work the authors present the first publicly available, open-source simulation model of the Third Generation Partnership Project (3GPP) Release 14 Cellular Vehicle to Everything (C-V2X) standard. The model specifically addresses Mode 4, in which vehicles communicate directly in peer to peer mode using a distributed resource allocation algorithm. Importantly, the model is fully compliant with existing service and application layers already defined by the automotive and standards communities, providing a fully standardised cross-layer communication model spanning application layer based on the ITS-G5 standard to physical layer from the new C-V2X Mode 4 standard. This enables enhancements of the lower layers to support future vehicular use cases.
Keywords: Networks LTE-V C-V2X 3GPP 4G Sensing Based Semi-Persistent Scheduling (SB-SPS)
Click here to view presentation
21/Aug/2019 12:00-13:00, WGB - G09, Technical piece of work
Eduardo Vyhmeister, Post-doc, Insight@UCC
Title: A combined photovoltaic and novel renewable energy system
Abstract: The productivity of the mining industry in Chile, currently the main driver of Chilean economy, is closely tied to foreign demand for ores. Ore-processing is known for involving energy-intensive processes, such as comminution, concentrating and cathodic processes. As mining activities take place in the arid north of Chile, they are affected by water scarcity. Water shortage has forced the industry to pump desalinated seawater up to mining sites over 2000 m above sea level, further increasing electricity consumption. Given these energy needs, and the fact that electrical energy supply in the north of Chile is based on fossil fuels, it is important to consider the use of renewable energies as environment-friendly and economic alternatives. The aim of this work is to evaluate, by an optimized techno-economic analysis, the use of photovoltaic and a novel wind-based technology to supply at least 10% of the current and the predicted electrical energy requirements of the mining industry in the Antofagasta region. A combination of an optimization problem and technical evaluation was performed using Matlab to obtain the optimal number of solar and wind-based technology units in a case study. Total energy generation from a novel wind-based technology unit is 67,616 MWh/y, corresponding to 14.45$10 6 Nm 3 (1298 t) of hydrogen and 7.41$10 6 Nm 3 (10,323 t) of oxygen after electrolytical transformation. Considering a 65% efficiency of the combined cycle fed with hydrogen and oxygen, 28,133 MWh/y of electrical energy would be obtained. For the cases studied the cost of energy from the combined system was estimated to be between 0.255 US$/kWh and 0.273 US$/kWh, slightly higher than the average energy regional cost. According to the analysis, the renewable energy system could be a sustainable alternative to supply economic green energy to the mining industry in Chile.
Keywords: AI Optimization Energy Hydrogen Electrolysis
Click here to view presentation
14/Aug/2019 12:00-13:00, WGB - G09, Workshop rehearsal
Andrea Visentin, Ph.D Student, Insight@UCC
Title: Optimal (R,s,S) policy for inventory lot sizing problem with stochastic non-stationary demand
Abstract: We propose an efficient branch-and-bound approach for computing the optimal (R, s, S) policy parameters under non-stationary stochastic demand. In this policy, the inventory position is raised to an order-up-to-level at a review instant when the inventory position is at or below reorder-levels. To the best of our knowledge, no method for computing the optimal policy parameters has been reported in the literature.
Keywords: AI Optimization Inventory Control Stochastic Demand Dynamic Programming
Click here to view presentation
31/July/2019 12:00-13:00, WGB - G01, Post-viva presentation
Begüm Genç, Post-doc, Confirm@UCC
Title: An approach to robustness in matching problems under ordinal preferences
Abstract: Matching problems under ordinal preferences is a popular field. Some of the well-known matching problems can be listed as Stable Marriage (SM), Stable Roommates (SR), and Kidney exchange problems. There are many variations of these problems being studied under different optimality criteria. In this work, considering the dynamism of the real-world, we propose using a novel 'robustness' notion for these problems. Our proposed notion that captures both robustness and stability constraints is called (a,b)-supermatches. We study (a,b)-supermatches for the SM and the SR. We present some theoretical findings related to the mathematical complexity as well as some empirical results. This talk mainly focuses on motivating the problem, discussing some real world examples, and presenting the outlines of the study.
Keywords: AI Optimization Theoretical Complexity Partially Ordered Sets Meta-heuristic Methods Matching Problems Robustness
Click here to view presentation
17/July/2019 12:00-13:00, WGB - G01, Post-viva presentation Mesut Kaya, Ph.D Student, Insight@UCC Title: Subprofile Aware Diversification of Recommendations
Abstract: Intent aware diversification methods formulate diversity in terms of coverage and relevance of aspects. The aspects are most commonly defined in terms of item features. We present a new form of intent aware diversification, which we call SPAD (Subprofile Aware Diversification). In SPAD and its variants, the aspects are not item features; they are subprofiles of the user’s profile. Therefore, they are useful even in domains where item features are not available or are of low quality. On several datasets, SPAD and its variants increase both relevance and diversity for even more configurations than other approaches to diversification.
Keywords: AI Recommender Systems Diversity Subprofiles

School of Computer Science and Information Technology

Scoil na Ríomheolaíochta agus na Teicneolaíochta Faisnéise

School of Computer Science and Information Technology, Western Gateway Building, University College Cork, Western Road, Cork, Ireland


Contact the School

csoffice@cs.ucc.ie

Contact the Seminar Team

seminarseries[at]cs.ucc.ie







The School of Computer Science and Information Technology Seminar Series is organised by a small team of Computer Science Researchers.