Personalisation in Cyber-Physical-Social Systems: A Multi-stakeholder aware Recommendation and Guidance

Abstract

The evolution of smart devices has led to the transformation of many physical spaces to the so-called smart environments collectively termed as Cyber-Physical-Social System (CPSS). In CPSS users co-exist with different stakeholders influencing each other while being influenced by different environmental factors. Additionally, these environments often have their own desired goals and corresponding set of rules in place expecting people to behave in certain ways. Hence, in such settings classical approaches to personalisation which solely optimise for user satisfaction are often encumbered by competing objectives and environmental constraints which are yet to be addressed jointly. In this work we set out to (i) formalise the general problem of personalisation in CPSS from a multi-stakeholder perspective taking into account the full environmental complexity, (ii) extend the general formalisation to the case of exhibition areas and propose a personalised Multi-stakeholder aware Recommendation and Guidance method on a case study of National Gallery, London.

The Elements of Visual Art Recommendation: Learning Latent Semantic Representations of Paintings

“…Finally, we note that our application asked participants to rate one painting randomly selected from each of the nine categories of our dataset. This resulted in a 9-dimensional preference elicitation vector with associated weights, which is perhaps small, considering that previous work asked participants to rate up to 80 paintings [73]. However, we have not observed substantial overlaps in the rankings produced by each RecSys engine, which indicates that each participant received truly personalized recommendations.…”

“…Overall, previous works showed that visual features tend to perform better than textual metadata [49,73] and hence they argued for not considering text-based information in VA RecSys. In addition, it has not been explored yet whether hybrid approaches may yield better performance on VA recommendation tasks.…”

“…Recent works by Yilma et al [72,73] proposed a VA recommendation approach that leveraged topic modeling techniques from textual descriptions of paintings and performed a comparative study against visual features automatically extracted using DNNs. Their study demonstrated the potential of learning features from textbased data, especially when it comes to explaining the recommendations to the user.…”

“…Thus, to enhance the perceived utility of RecSys, it is of paramount importance to emphasize visitors' quality of experience. In this context, Visual Art (VA) recommendation is among the areas that has recently gained momentum [73]. Nevertheless, contrary to other application areas of RecSys where personalised content is delivered to users such as movies, music, news, etc., the domain of VA recommendation has not yet been sufficiently explored.…”

“…Alternatively, Yilma et al [73] proposed an approach to learn latent visual and textual features from paintings. Their study indicated that recommendations derived from textual features compare favorably with visual ones.…”