Machine Learning Methods for Personalised, Abstractive Summarisation of Consumer-Generated Media

The success of Web 2.0 and CGM is based on tapping into the social nature of human interactions, by making it possible for people to voice their opinion, become part of a virtual community and collaborate remotely. If we take micro-blogging as an example, the growth in Twitter visits between 2008 and 2009 was over 1,000% and it is projected that by 2010 around 10% of all internet users will be on Twitter. This unprecedented rise in the volume and importance of online content has resulted in companies and individuals spending ever increasing amounts of time trying to keep up with relevant CGM. It is estimated that 700 person hours per year is the absolute minimum that companies and public services need to spend on CGM monitoring, online user engagement, and discovery of new information. This fellowship is about helping people to cope with the resulting information overload, through automatic methods that are capable of adapting to individual's information seeking goals and summarising briefly the relevant media and thus supporting information interpretation and decision making. Automatic text summarisation is key to our goal and consists of compressing the meaning of text documents while preserving the relevant information contained within them. While there has been a lot of research on well-authored texts such as news, summarisation of social media is still in its infancy, with research focused on product reviews. A key experimental finding has been that due to the characteristics of social media (product reviews in particular) it is better first to abstract the relevant information from the different documents and sites and then to use natural language generation to create a fluent text based on this information.In this fellowship I will investigate and evaluate new machine learning methods for personalised, abstractive multi-document summarisation across different social media. For example, diachronic summaries that combine Twitter posts, blog articles, and Facebook wall messages on a given topic. In contrast to previous work, we will pursue an inter-disciplinary approach, which will help us study the social dimension of CGM summarisation and establish actual user needs. The second research challenge is that the algorithms need to be robust in the face of this noisy, jargon-full and dynamic content, as well as needing models capable of representing the contradictory and strongly temporal nature of CGM. A key novel contribution of our work is personalising the summaries, based on a model of user interests, goals, and social context. Issues such as trustworthiness, privacy, and online communities (with their hubs and authorities) will also play an important role. The fourth research challenge is to generate personalised abstractive summaries that can help users with sensemaking and content interpretation. An exciting element of my research will be in studying the different kinds of summaries that are useful for a variety of real users (companies, journalists, and the general public) through multi-disciplinary collaborations with the Press Association, British Telecom, the Oxford Internet Institute, and Sheffield's Department of Journalism. A key project deliverable will be a publicly available browser plugin that provides easy access to the automatically generated summaries. This will allow me to evaluate the project results with real users, on a large scale. It will also provide a new evaluation challenge for the Natural Language Generation community, as researchers will be able to compare their summarisers against those delivered by our open-source algorithms. Last but not least, the fellowship covers not only foundational multi-disciplinary research but it also tests the results in several Digital Economy pilot experiments involving commercial partners (The Press Association, British Telecom, Fizzback).

Since consumer-generated media have revolutionised our personal lives, the economy, and society as whole, this research has wide-ranging implications and relevance. The five year duration of this fellowship will give us the necessary flexibility, time and resources to undertake dissemination, exploitation, and training activities aimed at the multiple beneficiaries detailed below. Firstly, in order to achieve maximum impact, project results will be made open-source, which will include both computational algorithms and the data collections on which they were tested and evaluated. Having open-source results is needed not only because we aim to create an active research community around the project results, but also since we want to promote use by non-governmental organisations, educational institutions, and other non-profit users. Another important project result will be a free tool to help users become aware of the dangers of social media and allow them to monitor continuously what is being disclosed about them online (by themselves, family, and friends). In terms of economic impact, we will target: (i) the areas used as Digital Economy pilots, namely digital journalism, voice-of-the-customer service providers, and online brand, product, and reputation management; (ii) companies providing internet privacy and security services; (iii) companies providing product comparison services to consumers. Major beneficiaries are the Digital Economy sectors identified above. In addition to the project partners (BT, PA, Fizzback, and Nokia), through UK and international research projects the PI has built successful industrial collaborations with other large companies (Yahoo, Atos, Dassault Aviation, Elsevier, MPS Bank, Creditreform, NetInfo) and SMEs (Garlik, Innovantage, Ontotext, Matrixware, Mondeca, Ontoprise, ISOCO), where our previous research showed the potential of intelligent summarisation technology for knowledge management and business intelligence. Further knowledge transfer opportunities arise through the Sheffield University connections to the digital and new media industries in the Sheffield city region, which are growing at a faster rate than anywhere else in the UK in terms of specialist companies and new jobs. A unique opportunity arises also from the 100m South Yorkshire Digital Region project, which will pilot the Next Generation Broadband and thus provide the required infrastructure for advanced digital economy applications. The research proposed here also has significant societal relevance. The first dimension is user privacy and trust, which are an important element of our work. We plan to promote the new technology to users and companies with identity theft products (e.g. Garlik, with whom we have worked in the past), in order to help people with tracking personal data divulged on public web sites and social networks. This research is relevant to policies such as: Digital Britain, EU's e2010, Global Focus on ICT in Development. We will reach out towards policy makers through the new UK Ethical ICT and young people stakeholder group and through Sheffield's Digital World outreach activities. Last, but not least, this fellowship will impact significantly the research careers of all team members. It will provide the PI with the unique opportunity to realise a step-change in her career, in order to join the 7% of female grade A staff in engineering and technology. The inter-disciplinary nature of our research team will help researchers to gain knowledge of fields complementary to their primary expertise and to develop new cross-disciplinary research skills. The PhD students will gain transferable research and presentation skills and be prepared for post-doctoral positions in academic or applied research.