Amparo Elizabeth Cano

Harnessing linked knowledge sources for topic classification in social media

Topic classification (TC) of short text messages offers an effective and fast way to reveal events happening around the world ranging from those related to Disaster (e.g. Sandy hurricane) to those related to Violence (e.g. Egypt revolution). Previous approaches to TC have mostly focused on exploiting individual knowledge sources (KS) (e.g. DBpedia or Freebase) without considering the graph structures that surround concepts present in KSs when detecting the topics of Tweets. In this paper we introduce a novel approach for harnessing such graph structures from multiple linked KSs, by: (i) building a conceptual representation of the KSs, (ii) leveraging contextual information about concepts by exploiting semantic concept graphs, and (iii) providing a principled way for the combination of KSs. Experiments evaluating our TC classifier in the context of Violence detection (VD) and Emergency Responses (ER) show promising results that significantly outperform various baseline models including an approach using a single KS without linked data and an approach using only Tweets.

Detecting child grooming behaviour patterns on social media

Online paedophile activity in social media has become a major concern in society as Internet access is easily available to a broader younger population. One common form of online child exploitation is child grooming, where adults and minors exchange sexual text and media via social media platforms. Such behaviour involves a number of stages performed by a predator (adult) with the final goal of approaching a victim (minor) in person. This paper presents a study of such online grooming stages from a machine learning perspective. We propose to characterise such stages by a series of features covering sentiment polarity, content, and psycho-linguistic and discourse patterns. Our experiments with online chatroom conversations show good results in automatically classifying chatlines into various grooming stages. Such a deeper understanding and tracking of predatory behaviour is vital for building robust systems for detecting grooming conversations and potential predators on social media.

Online Sentiment and Topic Dynamics Tracking over the Streaming Data

We propose a dynamic joint sentiment-topic model (dJST) which is able to effectively track sentiment and topic dynamics over the streaming data. Both topic and sentiment dynamics are captured by assuming that the current sentiment-topic specific word distributions are generated according to the word distributions at previous epochs. We study three different ways of accounting for such dependency information, (1) Sliding window where the current sentiment-topic-word distributions are dependent on the previous sentiment-topic specific word distributions in the last S epochs; (2) Skip model where history sentiment-topic-word distributions are considered by skipping some epochs in between; and (3) Multiscale model where previous long-and short-timescale distributions are taken into consideration. We derive efficient online inference procedures to sequentially update the model with newly arrived data and show the effectiveness of our proposed model on the Mozilla add-on reviews crawled between 2007 and 2011.

Stretching the Life of Twitter Classifiers with Time-Stamped Semantic Graphs

Social media has become an effective channel for communicating both trends and public opinion on current events. However the automatic topic classification of social media content pose various challenges. Topic classification is a common technique used for automatically capturing themes that emerge from social media streams. However, such techniques are sensitive to the evolution of topics when new event-dependent vocabularies start to emerge (e.g., Crimea becoming relevant to War_Conflict during the Ukraine crisis in 2014). Therefore, traditional supervised classification methods which rely on labelled data could rapidly become outdated. In this paper we propose a novel transfer learning approach to address the classification task of new data when the only available labelled data belong to a previous epoch. This approach relies on the incorporation of knowledge from DBpedia graphs. Our findings show promising results in understanding how features age, and how semantic features can support the evolution of topic classifiers.

Topica – Profiling Locations through Social Streams

This paper presents work in interlinking social stream information with geographical spaces through the use of Linked Data technologies. The paper focuses on filtering, enriching, structuring and interlinking microposts of localised (i.e. geo-tagged) social streams (a.k.a localised forums) to profile geographical areas (e.g., cities, countries). For this purpose, we enriched social streams extracted from Twitter, Facebook and TripAdvisor and structured them into well-known vocabularies and data models, such as SIOC and SKOS. To integrate this information into a location profile we introduce the linkedPOI ontology. The linkedPOI ontology captures and leverages DBpedia categories to derive concepts which profile a geographic space.

Understanding web documents using semantic overlays

The Ozone Browser is a platform independent tool that enables users to visually augment the knowledge presented in a web document in an unobtrusive way. This tool supports the user comprehension of Web documents through the use of Semantic Overlays. This tool uses linked data and lightweight semantics for getting relevant information within a document. The current implementation uses a JavaScript bookmarklet.

Semantic Topic Compass --- Classification Based on Unsupervised Feature Ambiguity Gradation

Characterising social media topics often requires new features to be continuously taken into account, and thus increasing the need for classifier retraining. One challenging aspect is the emergence of ambiguous features, which can affect classification performance. In this paper we investigate the impact of the use of ambiguous features in a topic classification task, and introduce the Semantic Topic Compass STC framework, which characterises ambiguity in a topics feature space. STC makes use of topic priors derived from structured knowledge sources to facilitate the semantic feature grading of a topic. Our findings demonstrate the proposed framework offers competitive boosts in performance across all datasets.

Representing, proving and sharing trustworthiness of web resources using veracity

The World Wide Web has evolved into a distributed network of web applications facilitating the publication of information on a large scale. Judging whether such information can be trusted is a difficult task for humans, often leading to blind trust. In this paper we present a model and the corresponding veracity ontology which allows trust to be placed in web content by web agents. Our approach differs from current work by allowing the trustworthiness of web content to be securely distributed across arbitrary domains and asserted through the provision of machine-readable proofs (i.e. by citing another piece of information, or stating the credentials of the user/agent). We provide a detailed scenario as motivation for our work and demonstrate how the ontology can be used.