HERMES: Trustworthy Multi-Objective and Multi-Stakeholder Recommenders

The HERMES project aims at researching and proposing state-of-the-art recommender system methods for Trustworthy Multi-objective and multi-stakeholder (MOS) environments. It will particularly focus on the domains of e-commerce and social networks. Both domains are highly similar from the recommender system perspective, as they share the same three elements – so-called stakeholders’ groups:
  • recommender system provider, 
  • recommended content consumers, 
  • and recommended content producers. 

In the MOS environments, a recommender system has to reflect, combine and solve often contrasting requirements and expectations from dozens of stakeholders. In fact, the problem is even more complicated, as each of the stakeholders can simultaneously have various objectives. For instance, the recommender system provider can optimize for customer long-term, dwell time, diversity, or click-through rate. This creates a diverse and complicated environment with several stakeholders and their objectives in which a recommender system operates.  Because of its nature each of the stakeholders operates with a different amount of information, which creates distrust of the specific stakeholders (e.g., customers). 

The project objectives advance the current state-of-the-art recommender systems. HERMES addresses the pressing basic research questions, which have, however, great applicability in industry applications. It will strengthen the existing research collaboration with other research teams across Europe. The research topic of the project also integrates the European approach to artificial intelligence, contributing to reducing the negative impacts of AI on society. The project will contribute to the state-of-the-art as follows:

  • Operationalization of the trustworthiness in the context of the recommender system and MOS environment
  • Self-assessment and checklist for trustworthy MOS recommender systems
  • Proposing novel trustworthy methods for MOS recommender systems (with a focus on e-commerce and social network characteristics)
Michal_Kompan

The project allows us to address hot topics in the recommender systems, namely how to create a trustworthy recommender with respect to several stakeholders and their objectives. We will also explore ethical aspects of such approaches, which will have a wider societal impact.

Michal Kompan

Chief Research Officer & Senior researcher, KInIT

Related content

Project outcomes (Deliverables):

Funded by the EU NextGenerationEU through the Recovery and Resilience Plan for Slovakia under the project No. 09I03-03-V04-00336.