Abstract-The advent of online social media and the growing popularity of sensor-equipped mobile devices have created a vast landscape of location-aware applications and services. This goldmine of data, including temporal and spatial information of unprecedented granularity, can help researchers gain insights into the behavioural patterns of people at a global scale. Here we analyse the textual content of millions of comments published alongside Foursquare user check-ins. For this, we extend a standard topic modelling approach so that it explicitly takes into account geographic and temporal side information. The framework is applied to Foursquare data and used to detect the dominant topics in the neighbourhoods of a city. In particular, we present the most prominent topics discussed by Foursquare users in New York, London, Chicago and San Francisco. We characterize the topics' spatial coverage and temporal evolution, and we also highlight some cultural idiosyncrasies. Finally, we evaluate the novel spatio-temporal topic model quantitatively. We believe that our model may be a useful tool for social scientists and application developers.

Abstract. This paper addresses a central sub-task of timeline creation from historical Wikipedia articles: learning from text which of the person names in a textual article should appear in a timeline on the same topic. We first process hundreds of timelines written by human experts and related Wikipedia articles to construct a corpus that can be used to evaluate systems that create history timelines from text documents. We then use a set of features to train a classifier that predicts the most important person names, resulting in a clear improvement over a competitive baseline.

Timeline generation is a summarisation task which transforms a narrative, roughly chronological input text into a set of timestamped summary sentences, each expressing an atomic historical event. We present a methodology for evaluating systems which create such timelines, based on a novel corpus consisting of 36 humancreated timelines. Our evaluation relies on deep semantic units which we call historical content units. An advantage of our approach is that it does not require human annotation of new system summaries.

No abstract

People commonly need to purchase things in person, from large garden supplies to home decor. Although modern search systems are very effective at finding online products, little research attention has been paid to helping users find places that sell a specific product offline. For instance, users searching for an apron are not typically directed to a nearby kitchen store by a standard search engine. In this paper, we investigate "where can I buy"-style queries related to in-person purchases of products and services. Answering these queries is challenging since little is known about the range of products sold in many stores, especially those which are smaller in size. To better understand this class of queries, we first present an in-depth analysis of typical offline purchase needs as observed by a major search engine, producing an ontology of such needs. We then propose ranking features for this new problem, and learn a ranking function that returns stores most likely to sell a queried item or service, even if there is very little information available online about some of the stores. Our final contribution is a new evaluation framework that combines distance with store relevance in measuring the effectiveness of such a search system. We evaluate our method using this approach and show that it outperforms a modern web search engine.