Benjamin Heitmann

ActiveRDF: Embedding Semantic Web data into object-oriented languages

Semantic Web applications share a large portion of development effort with database-driven Web applications. Existing approaches for development of these database-driven applications cannot be directly applied to Semantic Web data due to differences in the underlying data model. We develop a mapping approach that embeds Semantic Web data into object-oriented languages and thereby enables reuse of existing Web application frameworks. We analyse the relation between the Semantic Web and the Web, and survey the typical data access patterns in Semantic Web applications. We discuss the mismatch between object-oriented programming languages and Semantic Web data, for example in the semantics of class membership, inheritance relations, and object conformance to schemas. We present ActiveRDF, an object-oriented API for managing RDF data that offers full manipulation and querying of RDF data, does not rely on a schema and fully conforms to RDF(S) semantics. ActiveRDF can be used with different RDF data stores: adapters have been implemented to generic SPARQL endpoints, Sesame, Jena, Redland and YARS and new adapters can be added easily. We demonstrate the usage of ActiveRDF and its integration with the popular Ruby on Rails framework which enables rapid development of Semantic Web applications.

A Flexible Integration Framework for Semantic Web 2.0 Applications

The Semantic Web application framework extends Ruby on Rails to enable rapid development of integrated Semantic Web mash-ups. Web applications are mostly database driven. Developers design a database schema and then construct the application logic (which generates Web pages for user interaction) on top of the schema. These applications are centralized and rely on their own relational database, limiting the possibilities for data integration. Mash-ups (often called Web 2.0 applications) are an emerging Web development paradigm that combines functionality from different Web applications.

An open framework for multi-source, cross-domain personalisation with semantic interest graphs

Cross-domain recommendations are currently available in closed, proprietary social networking ecosystems such as Facebook, Twitter and Google+. I propose an open framework as an alternative, which enables cross-domain recommendations with domain-agnostic user profiles modeled as semantic interest graphs. This novel framework covers all parts of a recommender system. It includes an architecture for privacy-enabled profile exchange, a distributed and domain-agnostic user model and a cross-domain recommendation algorithm. This enables users to receive recommendations for a target domain (e.g. food) based on any kind of previous interests.

An Empirically Grounded Conceptual Architecture for Applications on the Web of Data

We present a component-based, conceptual architecture for Semantic Web applications. It describes the high-level functionality that substantially differentiates Resource Description Framework (RDF)-supported applications from database-driven applications. We provide a strong empirical grounding for this architecture through a survey of Semantic Web applications over most of the past decade. Our empirical approach allows us to describe the current state of the art for the development and deployment of applications on the Web of Data. In addition, we determine how far the adoption of signature research topics of the Semantic Web, such as, data reuse, data integration, and reasoning, has progressed. We, then, discuss the main implementation challenges that developers using Semantic technologies are facing, as observed in the survey. We build on this in order to suggest future approaches to facilitate the standardization of components and the development of software engineering tools to increase the uptake of the Web of Data.

SemStim at the LOD-RecSys 2014 Challenge

SemStim is a graph-based recommendation algorithm which is based on Spreading Activation and adds targeted activation and duration constraints. SemStim is not aected by data sparsity, the cold-start prob- lem or data quality issues beyond the linking of items to DBpedia. The overall results show that the performance of SemStim for the diversity task of the challenge is comparable to the other participants, as it took 3rd place out of 12 participants with 0.0413 F1@20 and 0.476 ILD@20. In addition, as SemStim has been designed for the requirements of cross-domain rec- ommendations with dierent target and source domains, this shows that SemStim can also provide competitive single-domain recommendations.

XploDiv: Diversification Approach for Recommender Systems

Recommender Systems have emerged to guide users in the task of efficiently browsing/exploring a large product space, helping users to quickly identify interesting products. However, suggestions generated with traditional Recommender Systems usually do not produce diverse results, though it has been argued that diversity is a desirable feature. The study of diversity aware Recommender Systems has become an important research challenge in recent years, drawing inspiration from diversification solutions for Information Retrieval. However, we argue it is not enough to adapt Information Retrieval techniques towards Recommender Systems, as they do not place the necessary importance to factors such as serendipity, novelty and discovery which are imperative to Recommender Systems. In this report, we propose a diversification technique for Recommender Systems that generates a diversified list of results which not only balances the trade-off between quality (in terms of accuracy) and diversity, but also considers the trade-off between exploitation of the user profile and exploration of novel products. Our experimental evaluation, composed of both qualitative and quantitative tests, shows that the proposed approach has comparable results to state of the art approaches. Moreover, through control parameters, our approach can be tuned towards more explorative or exploitative recommendations.

SemStim: Exploiting Knowledge Graphs for Cross-Domain Recommendation

In this paper we introduce SemStim, an unsupervised graph-based algorithm that addresses the cross-domain recommendation task. In this task, preferences from one conceptual domain (e.g. movies) are used to recommend items belonging to another domain (e.g. music). SemStim exploits the semantic links found in a knowledge graph (e.g. DBpedia), to connect domains and thus generate recommendations. As a key benefit, our algorithm does not require (1) ratings in the target domain, thus mitigating the cold-start problem and (2) overlap between users or items from the source and target domains. In contrast, current state-of-the-art personalisation approaches either have an inherent limitation to one domain or require rating data in the source and target domains. We evaluate SemStim by comparing its accuracy to state-of-the-art algorithms for the top-k recommendation task, for both single-domain and cross-domain recommendations. We show that SemStim enables cross-domain recommendation, and that in addition, it has a significantly better accuracy than the baseline algorithms.

Towards Near Real-Time Social Recommendations for the Enterprise

The widespread use of social platforms in contemporary organizations leads to the generation of large amounts of content shared through various social tools. This information is distributed and often unstructured, making it difficult to fully exploit its value in an enterprise context. While Semantic Web technologies allow for publishing meaningful and structured data, major challenges include: (1) real-time integration of distributed social data, and (2) content personalization to identify relevant pieces of information and present them to users to limit the information overload. We propose to combine Semantic Web technologies with standardized transport protocols, such as XMPP, to provide an efficient and open source layer for aggregation of distributed social data in an enterprise. In addition, we propose a personalisation approach, which is able to provide filtered and personalised access on top of such distributed social data.

Using social media data for online television recommendation services at RTÉ Ireland

Raidió Teilifís Éireann (RTÉ) is the public service television and radio broadcaster in Ireland. Through on demand video services, RTÉ allows their users to catch up on television broadcasts via the RTÉ Player. The company interacts with their users by means of social media platforms such as Twitter, Facebook and YouTube. Aiming to improve user engagement and in order to deal with the potential information overload caused by a broad catalogue such as RTÉ’s, a recommendation service would be a desirable feature for the RTÉ Player. However, the distinctive requirements of this use case, such as (a) lack of ratings, (b) lack of user sessions data, and (c) limited lifespan of items (i.e., available videos), keep us from applying traditional recommendation approaches. Yet, we believe that immersed within social media data, there is valuable information about user viewing preferences, and that this information can be used as input for Recommendation Services. This paper examines the use of social media data for personalization in the RTÉ industrial use case, giving special focus to the use of Twitter data for Recommendation Systems.