Session details: XperienceWeb'12 Workshop 2

We are pleased to present the proceedings of the XperienceWeb 2012 workshop, which was held as part of the World Wide Web 2012 Conference in Lyon, April 2012. This workshop has evolved out of the webCBR workshop series, which primarily promotes the application of case-based reasoning (CBR) to experiential content on the web. Clearly such content is on the increase with advances in web technology resulting in vast amounts of user generated web content in the form of blogs, tweets, forums and opinion sites. Whilst such content can be treated as documents we take the view that they are more appropriately viewed as a rich source of untapped experience data, a valuable asset that can be used to generate Web experience bases through CBR technology. The key idea in CBR involves the reuse of similar experiences to resolve new problems. Such reasoning is relevant to the web because increasingly people search and browse other people's experiences on travel, medicine, retail, entertainment, etc., for their informational and problem solving needs. Typically this vast repository of untapped knowledge and experience exists in unstructured text and multimedia forms. These rich informational snippets generally describe a particular problem situation often combined with a reflective narrative consisting of decisions made, lessons learnt and opinions expressed. Such content forms a rich source of untapped experience data - an integral resource for problem solving with CBR. For example, if one wants to identify web content that helps them to achieve a particular task (say software configuration and installation), then it is likely that there will be many related user posts describing experiences on similar tasks and crucially include suitable solutions to address such a task. We therefore envisage a new form of user need, one that possibly embodies knowledge, meaning and understanding as opposed to chunks of keywords for queries and snippets. Addressing this need calls for novel extraction, representation, retrieval and assembly strategies from existing experiential content on the web. The focus of this workshop was to provide a forum for the discussion of trends, research issues and practical experiences on the role of tools and technologies in reasoning with web-related experiential content. Six papers were accepted for the workshop, covering a wide range of web experience data. Adekyanju et al. present an approach for combining knowledge from web data and search logs for query recommendation. Kato et al. propose a method for extracting onomatopoeia from restaurant reviews and using these in a restaurant recommender. Schumacher et al. compare two methods for extracting procedural or 'how-to' knowledge from the web and Zarka et al. look at the potential for experiential approaches to recommendation to enable a more context-aware method of generating useful video recommendations, particularly in relation to temporal context. Two of the accepted papers are position papers. The question of how user-generated content can be mined successfully for solutions to specific problems is discussed by Sauer & Roth-Berghofer while Gorg et al. consider the problem of constructing and managing personal workflows over a social network.