Tiphaine Dalmas

Answer comparison in automated question answering

Abstract In the context of Question Answering (QA) on free text, we assess the value of answer comparison and information fusion in handling multiple answers. We report improvements in answer re-ranking using fusion on a set of location questions and show the advantages of considering candidates as allies rather than competitors. We conclude with some observations about answer modeling and evaluation methodology, arising from a more recent experiment with a larger set of questions and a greater diversity of question types and candidates.

Learning foci for Question Answering over Topic Maps

This paper introduces the concepts of asking point and expected answer type as variations of the question focus. They are of particular importance for QA over semistructured data, as represented by Topic Maps, OWL or custom XML formats. We describe an approach to the identification of the question focus from questions asked to a Question Answering system over Topic Maps by extracting the asking point and falling back to the expected answer type when necessary. We use known machine learning techniques for expected answer type extraction and we implement a novel approach to the asking point extraction. We also provide a mathematical model to predict the performance of the system.

On the Implementation of HealthAgents: Agent Based Brain Tumour Diagnosis

This paper introduces HealthAgents, an EC-funded research project to improve the classification of brain tumours through multi-agent decision support over a secure and distributed network of local databases or Data Marts. HealthAgents will not only develop new pattern recognition methods for distributed classification and analysis of in vivo MRS and ex vivo/in vitro HRMAS and DNA data, but also define a method to assess the quality and usability of a new candidate local database containing a set of new cases, based on a compatibility score. Using its Multi-Agent architecture, HealthAgents intends to apply cutting-edge agent technology to the Biomedical field and develop the HealthAgents network, a globally distributed information and knowledge repository for brain tumour diagnosis and prognosis.

Cross-lingual question answering using off-the-shelf machine translation

We show how to adapt an existing monolingual open-domain QA system to perform in a cross-lingual environment, using off-the-shelf machine translation software. In our experiments we use French and German as source language, and English as target language. For answering factoid questions, our system performs with an accuracy of 16% (German to English) and 20% (French to English), respectively. The loss of correctly answered questions caused by the MT component is estimated at 10% for French, and 15% for German. The accuracy of our system on correctly translated questions is 28% for German and 29% for French.

A dialogue based mobile virtual assistant for tourists: The SpaceBook Project

Abstract Ubiquitous mobile computing offers innovative approaches in the delivery of information that can facilitate free roaming of the city, informing and guiding the tourist as the city unfolds before them. However making frequent visual reference to mobile devices can be distracting, the user having to interact via a small screen thus disrupting the explorative experience. This research reports on an EU funded project, SpaceBook, that explored the utility of a hands-free, eyes-free virtual tour guide, that could answer questions through a spoken dialogue user interface and notify the user of interesting features in view while guiding the tourist to various destinations. Visibility modelling was carried out in real-time based on a LiDAR sourced digital surface model, fused with a variety of map and crowd sourced datasets (e.g. Ordnance Survey, OpenStreetMap, Flickr, Foursquare) to establish the most interesting landmarks visible from the users location at any given moment. A number of variations of the SpaceBook system were trialled in Edinburgh (Scotland). The research highlighted the pleasure derived from this novel form of interaction and revealed the complexity of prioritising route guidance instruction alongside identification, description and embellishment of landmark information – there being a delicate balance between the level of information ‘pushed’ to the user, and the users requests for further information. Among a number of challenges, were issues regarding the fidelity of spatial data and positioning information required for pedestrian based systems – the pedestrian having much greater freedom of movement than vehicles.