DSpace Community: CLARITY: Centre for Sensor Web Technologies

A multi-criteria evaluation of a user generated content based recommender system

Sun, 23 Oct 2011 00:00:00 GMT

Author: Garcia Esparza, Sandra; O'Mahony, Michael P.; Smyth, Barry

Abstract: The Social Web provides new and exciting sources of information that may be used by recommender systems as a complementary source of recommendation knowledge. For example, User-Generated Content, such as reviews, tags, comments, tweets etc. can provide a useful source of item information and user preference data, if a clear signal can be extracted from the inevitable noise that exists within these sources. In previous work we explored this idea, mining term-based recommendation knowledge from user reviews, to develop a recommender that compares favourably to conventional collaborative-filtering style techniques across a range of product types. However, this previous work focused solely on recommendation accuracy and it is now well accepted in the literature that accuracy alone tells just part of the recommendation story. For example, for many, the promise of recommender systems lies in their ability to surprise with novel recommendations for less popular items that users might otherwise miss. This makes for a riskier recommendation prospect, of course, but it could greatly enhance the practical value of recommender systems to end-users. In this paper we analyse our User-Generated Content (UGC) approach to recommendation using metrics such as novelty, diversity, and coverage and demonstrate superior performance, when compared to conventional user-based and item- based collaborative filtering techniques, while highlighting a number of interesting performance trade-offs.

Conference details: Presented at the 3rd Workshop on Recommender Systems and the Social Web (RSWEB-11), 5th ACM Conference on Recommender Systems, Chicago, IL, USA, 23-27 October 2011

Power to the people : exploring neighbourhood formations in social recommender systems

Thu, 01 Sep 2011 00:00:00 GMT

Author: Bourke, Steven; McCarthy, Kevin; Smyth, Barry

Abstract: The explosive growth of online social networks in recent times has presented a powerful source of information to be utilised in personalised recommendations. Unsurprisingly there has already been a large body of work completed in the recommender system field to incorporate this social in- formation into the recommendation process. In this paper we examine the practice of leveraging a user’s social graph in order to generate recommendations. Using various neighbourhood selection strategies, we examine the user satisfaction and the level of perceived trust in the recommendations received.

Conference details: Paper presented at RecSys-11, Chicago IL, USA, October 23-27, 2011

Models of web page reputation in social search

Sun, 09 Oct 2011 00:00:00 GMT

Author: McNally, Kevin; O'Mahony, Michael P.; Smyth, Barry

Abstract: To date web search has been a solitary experience for the end-user, despite the fact that recent studies highlight the potential for collaboration that is inherent in many search tasks and scenarios. As a result, researchers have begun to explore the potential for a more collaborative approach to web search, one in which the search actions of other users can influence the results returned. In this context, the expertise of other users plays an important role when it comes to ensuring the quality of recommendations that arise from their actions. The reputation of these users is important in collaborative and social search tasks, much as relevance is vital in conventional web search. In this paper we examine this concept of reputation in collaborative and social search contexts. We describe a number of different reputation models and evaluate them in the context of a particular social search service. Our results highlight the potential for reputation to improve the quality of recommendations that arise from the activities of other searchers.

Conference details: Paper presented at the Third IEEE International Conference on Social Computing (SocialCom2011), MIT, Boston, USA, 9-11 October 2011

Pervasive computing technologies for healthcare

Thu, 01 Sep 2011 00:00:00 GMT

Author: O'Grady, Michael J.; O'Hare, G. M. P. (Greg M. P.); O'Donoghue, John

Abstract: The conference series on Pervasive Computing Technologies for Healthcare is one of the leading fora for research dissemination in this space. In May 2011, the most recent event took place in Ireland. A brief overview of the conference is now presented.

Real-time recognition and profiling of appliances through a single electricity sensor

Tue, 01 Jun 2010 00:00:00 GMT

Author: Ruzzelli, Antonio G.; Nicolas, C.; Schoofs, Anthony; O'Hare, G. M. P. (Greg M. P.)

Abstract: Sensing, monitoring and actuating systems are expected to play a key role in reducing buildings overall energy consumption. Leveraging sensor systems to support energy efficiency in buildings poses novel research challenges in monitoring space usage, controlling devices, interfacing with smart energy meters and communicating with the energy grid. In the attempt of reducing electricity consumption in buildings, identifying individual sources of energy consumption is key to generate energy awareness and improve efficiency of available energy resources usage. Previous work studied several non-intrusive load monitoring techniques to classify appliances; however, the literature lacks of an comprehensive system that can be easily installed in existing buildings to empower users profiling, benchmarking and recognizing loads in real-time. This has been a major reason holding back the practice adoption of load monitoring techniques. In this paper we present RECAP: RECognition of electrical Appliances and Profiling in real-time. RECAP uses a single wireless energy monitoring sensor easily clipped to the main electrical unit. The energy monitoring unit transmits energy data wirelessly to a local machine for data processing and storage. The RECAP system consists of three parts: (1) Guiding the user for profiling electrical appliances within premises and generating a database of unique appliance signatures; (2) Using those signatures to train an artificial neural network that is then employed to recognize appliance activities (3) Providing a Load descriptor to allow peer appliance benchmarking. RECAP addresses the need of an integrated and intuitive tool to empower building owners with energy awareness. Enabling real-time appliance recognition is a stepping-stone towards reducing energy consumption and allowing a number of major applications including load-shifting techniques, energy expenditure breakdown per appliance, detection of power hungry and faulty appliances, and recognition of occupant activity. This paper describes the system design and performance evaluation in domestic environment.

Conference details: Paper presented at Sensor Mesh and Ad Hoc Communications and Networks (SECON), 2010 7th Annual IEEE Communications Society Conference, Boston, Massachusetts, 21-25 June, 2010

Evaluating user reputation in collaborative web search

Sun, 23 Oct 2011 00:00:00 GMT

Author: McNally, Kevin; O'Mahony, Michael P.; Smyth, Barry

Abstract: Often today’s recommender systems look to past user activity in order to influence future recommendations. In the case of social web search, employing collaborative recommendation techniques allows for personalization of search results. If recommendations arise from past user activity, the expertise of those users driving the recommendation process can play an important role when it comes to ensuring recommendation quality. Hence the reputation of users is important in collaborative and social search tasks, in addition to result relevance as traditionally considered in web search. In this paper we explore this concept of reputation; specifically, investigating how reputation can enhance the recommendation engine at the core of the HeyStaks social search utility. We evaluate a number of different reputation models in the context of the HeyStaks system, and demonstrate how incorporating reputation into the recommendation process can enhance the relevance of results recommended by HeyStaks.

Conference details: Presented at the 3rd Workshop on Recommender Systems and the Social Web (RSWEB-11), 5th ACM Conference on Recommender Systems, Chicago, IL, USA, 23-27 October 2011.

Recommending case bases : applications in social web search

Thu, 15 Sep 2011 00:00:00 GMT

Author: Saaya, Zurina; Smyth, Barry; Coyle, Maurice; Briggs, Peter

Abstract: For the main part, when it comes to questions of retrieval, the focus of CBR research has been on the retrieval of cases from a repository of experience knowledge or case base. In this paper we consider a complementary retrieval issue, namely the retrieval of case bases themselves in scenarios where experience may be distributed across multiple case repositories. We motivate this problem with reference to a deployed social web search service called HeyStaks, which is based on the availability of multiple repositories of shared search knowledge, known as staks, and which is fully integrated into mainstream search engines in order to provide a more collaborative search experience. We describe the case base retrieval problem in the context of HeyStaks, propose a number of case base retrieval strategies, and evaluate them using real-user data from recent deployments.

Conference details: Paper presented at the International Conference on Case Based Reasoning (ICCBR-11), Greenwich, London, UK, 12-15 September, 2011

SimpleFlow : enhancing gestural interaction with gesture prediction, abbreviation and autocompletion

Thu, 01 Sep 2011 00:00:00 GMT

Author: Bennett, Mike; McCarthy, Kevin; O'Modhrain, Sile; Smyth, Barry

Abstract: Gestural interfaces are now a familiar mode of user interaction and gestural input is an important part of the way that users can interact with such interfaces. However, entering gestures accurately and efficiently can be challenging. In this paper we present two styles of visual gesture autocompletion for 2D predictive gesture entry. Both styles enable users to abbreviate gestures. We experimentally evaluate and compare both styles of visual autocompletion against each other and against non-predictive gesture entry. The best perform- ing visual autocompletion is referred to as SimpleFlow. Our findings establish that users of SimpleFlow take significant advantage of gesture autocompletion by entering partial gestures rather than whole gestures. Compared to non- predictive gesture entry, users enter partial gestures that are 41% shorter than the complete gestures, while simultaneously improving the accuracy (+13%, from 68% to 81%) and speed (+10%) of their gesture input. The results provide insights into why SimpleFlow leads to significantly enhanced performance, while showing how predictive gestures with simple visual autocompletion impacts upon the gesture abbreviation, accuracy, speed and cognitive load of 2D predictive gesture entry.

Conference details: Paper presented at Human-Computer Interaction – INTERACT 2011, 13th IFIP TC 13 International Conference, Lisbon, Portugal, September 5-9, 2011

Further experiments in micro-blog categorization

Wed, 31 Aug 2011 00:00:00 GMT

Author: Garcia Esparza, Sandra; O'Mahony, Michael P.; Smyth, Barry

Abstract: Since the creation of Twitter in 2008, micro-blogging services have received a lot of attention among users who wish to share news items, opinions and information with friends and colleagues. However, these services typically provide for only limited organisation of content, with the main ranking criterion being post time with perhaps some basic message filtering accommodated. Given the substantial and increasing volume of posts that micro-blogging services attract, there is a clear need to assist users when it comes to effectively consuming this content. In this regard, categorisation offers one approach to organise content by grouping related messages together. In this paper we present a study in the recommendation of categories for short-form messages in order to provide for better search and message filtering. In particular, we present an index-based approach where real-time web data can be used as a source of knowledge for category recommendation. Further, we evaluate our approach on two different micro-blogging datasets and results show that micro-blog messages in sufficient quantities provide a useful recommendation signal for category recommendation.

Conference details: Paper presented at the 22nd Irish Conference on Artificial Intelligence and Cognitive Science (AICS 2011), University of Ulster, Northern Ireland, 31 August - 2 September, 2011

Lowering the bar for robotic development : driver generation for ubiquitous robotic systems

Wed, 31 Aug 2011 00:00:00 GMT

Author: Treanor, Jennifer; O'Hare, G. M. P. (Greg M. P.)

Abstract: Robotics has developed, technologically, to a level where it becomes a field of both interest and importance to other disciplines, either as a proof-of-concept or demonstrative tool, or else as the main focus for implementation of theories. This is particularly evident in the areas of computational and theoretical cognitive science where, despite this progress, robotics remains sufficiently inaccessible to non-specialists as to dissuade its use. This is due in no small part to the issue of code re-usability across differing hardware platforms and the lack of low-level support for developing suitable drivers for the main robotics development tools. To address this issue, this work presents ACorDE: Autonomous Control Development Environment. This development environment takes in data pertaining to the robotic platform and generates suitable driver and behavioural code in a standardised format.

Conference details: Paper presented at the 22nd Irish Conference on Artificial Intelligence and Cognitive Science (AICS 2011), University of Ulster, Northern Ireland, 31 August - 2 September, 2011

The pursuit of happiness : searching for worthy followees on twitter

Wed, 31 Aug 2011 00:00:00 GMT

Author: Hannon, John; McCarthy, Kevin; Smyth, Barry

Abstract: We are living in an age of information overload, where it can be difficult to define which information is relevant and important to the end user at a point in time. In this paper, we introduce a solution to apportioning this constant flow of information by going to the source of the content, namely the producers. This paper examines an application for searching for pertinent friends on the popular microblogging service, Twitter1 and our approach to curtail the cold start problem that new users of the service face. We introduce our search technology which is capable of finding the producers of wanted content and suggest connecting to them as followees on Twitter. We also prove the usefulness of this technology through the results of a live user experiment carried out on these cold start users.

Conference details: Paper presented at the 22nd Irish Conference on Artificial Intelligence and Cognitive Science (AICS 2011), University of Ulster, Northern Ireland, 31 August - 2 September, 2011

Recommending search experiences

Wed, 31 Aug 2011 00:00:00 GMT

Author: Saaya, Zurina; Smyth, Barry; Coyle, Maurice; Briggs, Peter

Abstract: In this paper we focus on a multi-case case-based reasoning system to support users during collaborative search tasks. In particular we describe how repositories of search experiences/knowledge can be recommended to users at search time. These recommendations are evaluated using real-world search data.

Conference details: Paper presented at the 22nd Irish Conference on Artificial Intelligence and Cognitive Science (AICS 2011), University of Ulster, Northern Ireland, 31 August - 2 September, 2011

Modeling user and result reputation in collaborative web search

Wed, 31 Aug 2011 00:00:00 GMT

Author: McNally, Kevin; O'Mahony, Michael P.; Smyth, Barry

Abstract: Employing collaborative recommendation techniques allows for personalization of search results in social web search. If recommendations arise from past user activity, the expertise of those users driving the recommendation process can play an important role when it comes to ensuring recommendation quality. Hence the reputation of users is important, in addition to result relevance as traditionally considered in web search. In this paper we explore this concept of reputation; specifically, investigating how reputation can enhance the recommendation engine at the core of the HeyStaks social search utility. We evaluate a number of different reputation models in the context of the HeyStaks system, and demonstrate how incorporating reputation into the recommendation process can enhance the relevance of results recommended by HeyStaks.

Conference details: Paper presented at the 22nd Irish Conference on Artificial Intelligence and Cognitive Science (AICS 2011), University of Ulster, Northern Ireland, 31 August - 2 September, 2011

Using social ties in group recommendation

Wed, 31 Aug 2011 00:00:00 GMT

Author: Bourke, Steven; McCarthy, Kevin; Smyth, Barry

Abstract: The social web is a mass of activity, petabytes of data are generated yearly. The social web has proven to be a great resource for new recommender system techniques and ideas. However it would appear that typically these techniques are not so social, as they only generate recommendations for a user acting alone. In this paper we take the social graph data and preference content (via Facebook) of 94 user study participants and generate social group recommendations for them and their friends. We evaluate how different aggregation policies perform in deciding the final group recommendation. Our findings show that in an offline evaluation an aggregation policy which takes into consideration social weighting outperforms other aggregation policies.

Conference details: Paper presented at the 22nd Irish Conference on Artificial Intelligence and Cognitive Science (AICS 2011), University of Ulster, Northern Ireland, 31 August - 2 September, 2011

Demo abstract : appliance load monitoring by power load disaggregation

Tue, 01 Jun 2010 00:00:00 GMT

Author: Schoofs, Anthony; Sintoni, Alex; O'Hare, G. M. P. (Greg M. P.); Ruzzelli, Antonio G.

Abstract: Appliance load monitoring systems are designed to disaggregate the power load of a building in order to estimate the nature of individual loads, providing a real-time fine-grained recognition of active appliances. Monitoring non-intrusively appliances’ contributions to a given load enables a wide range of applications, ranging from electricity bill decomposition to accurate electricity user profiling. This work demonstrates a real implementation of such appliance load monitoring system. An intuitive graphical user interface is proposed to drive the system setup for profiling appliances’ signatures and for visualising the monitoring output.

Conference details: Paper presented at the 2010 7th Annual IEEE Communications Society Conference on Sensor, Mesh and Ad Hoc Communications and Networks (SECON), 21-25 June, 2010, Boston, Massachusetts

A mobile gateway for remote interaction with wireless sensor networks

Thu, 09 Jun 2011 00:00:00 GMT

Author: Angove, Philip; O'Grady, Michael J.; Hayes, Jer; O'Flynn, Brendan; O'Hare, G. M. P. (Greg M. P.); Diamond, Dermot

Abstract: Wireless Sensor Networks (WSNs) almost invariably support a centralised network management model. Though the data gathering function is conducted remotely, such data is usually routed via data sinks to central servers for processing, storage, visualisation and interpretation. However, the issue of supporting remote access to WSNs and individual sensor nodes whilst in their physical environment has not been viewed as a priority. It is envisaged that this situation will change as WSNs proliferate in a range of domains, and the potential for supporting innovative revenue-generating services manifest themselves. As a step towards realising such access, a mobile gateway has been designed and implemented. This gateway supports Zigbee as this is the predominant protocol supported by WSNs. Furthermore, it also supports Bluetooth, thereby facilitating interaction with conventional mobile devices. The gateway is programmable according to the needs of arbitrary services and applications.

MiRA - mixed reality agents

Fri, 01 Apr 2011 00:00:00 GMT

Author: Holz, Thomas; Campbell, Abraham G.; O'Hare, G. M. P. (Greg M. P.); Stafford, John W.; Dragone, Mauro; Martin, A. (Alan)

Abstract: In recent years, an increasing number of Mixed Reality (MR) applications have been developed using agent technology—both for the underlying software and as an interface metaphor. However, no unifying field or theory currently exists that can act as a common frame of reference for these varied works. As a result, much duplication of research is evidenced in the literature. This paper seeks to fill this important gap by outlining ‘‘for the first time’’ a formal field of research that has hitherto gone unacknowledged, namely the field of Mixed Reality Agents (MiRAs), which are defined as agents embodied in a Mixed Reality environment. Based on this definition, a taxonomy is offered that classifies MiRAs along three axes: agency, based on the weak and strong notions outlined by Wooldridge and Jennings (1995); corporeal presence, which describes the degree of virtual or physical representation (body) of a MiRA; and interactive capacity, which characterises its ability to sense and act on the virtual and physical environment. Furthermore, this paper offers the first comprehensive survey of the state-of-the-art of MiRA research and places each project within the proposed taxonomy. Finally, common trends and future directions for MiRA research are discussed. By defining Mixed Reality Agents as a formal field, establishing a common taxonomy, and retrospectively placing existing MiRA projects within it, future researchers can effectively position their research within this landscape, thereby avoiding duplication and fostering reuse and interoperability.

PI : perceiver and interpreter of smart home datasets

Mon, 23 May 2011 00:00:00 GMT

Author: Ye, Juan; Stevenson, Graeme; Dobson, Simon; O'Grady, Michael J.; O'Hare, G. M. P. (Greg M. P.)

Abstract: Pervasive healthcare systems facilitate various aspects of research including sensor technology, software technology, artificial intelligence and human-computer interaction. Researchers can often benefit from access to real-world data sets against which to evaluate new approaches and algorithms. Whilst more than a dozen data sets are currently publicly available, their use of heterogeneous mark-up impedes easy and widespread use. We describe PI – the Perceiver and semantic Interpreter – which offers a workbench API for the querying, re-structuring and re-purposing of a range of diverse data formats currently in use. The use of a single API reduces cognitive overload, improves access, and supports integration of generic and domain-specific information within a common framework.

Conference details: Paper presented at Pervasive Health 2011, 5th International ICST Conference on Pervasive Computing Technologies for Healthcare, 23rd-26th May 2011 - Dublin, Ireland

Orange alerts : lessons from an outdoor case study.

Mon, 23 May 2011 00:00:00 GMT

Author: Wan, Jie; Byrne, Caroline; O'Hare, G. M. P. (Greg M. P.); O'Grady, Michael J.

Abstract: Ambient Assisted Living (AAL) is of particular relevance to those who may suffer from Alzheimer’s Disease or dementia, and, of course, their carers. The slow but progressive nature of the disease, together with its neurological nature, ultimately compromises the behavior and function of people who may be essentially healthy from a physical perspective. An illustration of this is the wandering behavior frequently found in people with dementia. In this paper, a novel AAL solution for caregivers, particularly tailored for Alzheimer’s patients who are in the early stage of the disease and exhibit unpredictable wandering behavior, is briefly described. Salient aspects of a user evaluation are presented, and some issues relevant to the practical design of AAL systems in dementia cases are identified.

Conference details: Workshop at 5th International ICST Conference on Pervasive Computing Technologies for Healthcare, 23rd-26th May 2011 - Dublin, Ireland

Generating power footprints without appliance interaction : an enabler for privacy intrusion

Mon, 27 Jun 2011 00:00:00 GMT

Author: Sintoni, Alex; Schoofs, Anthony; Doherty, A.; Smeaton, Alan F.; O'Hare, G. M. P. (Greg M. P.); Ruzzelli, Antonio G.

Abstract: Appliance load monitoring (ALM) systems are systems capable of monitoring appliances’ operation within a building using a single metering point. As such, they uncover information on occupants’ activities of daily living and subsequently an exploitable privacy leak. Related work has shown monitoring accuracies higher than 90% ̇ achieved by ALM systems, yet requiring interaction with appliances for system calibration. In the context of external privacy intrusion, ALM systems have the following obstacles for system calibration: (1) type and model of appliances inside the monitored building are entirely unknown; (2) appliances cannot be operated to record power footprints; and (3) ground truth data is not available to fine- tune algorithms. Within this work, we focus on monitoring those appliances from which we can infer occupants’ activities. Without appliance interaction, appliances’ profiling is realised via automated capture and analysis of shapes, steady-state durations, and occurrence patterns of power loads. Such automated process produces unique power footprints, and naming is realised using heuristics and known characteristics of typical home equipment. Data recorded within a kitchen area and one home illustrates the various processing steps, from data acquisition to power footprint naming.

Conference details: Paper presented at the 1st HOBNET Workshop on IPv6 Sensor Networking for Smart/Green Buildings, at the 7th IEEE International Conference on Distributed Computing in Sensor Systems DCOSS '11, June 27 - 29, 2011, Barcelona, Spain