Contextual Scalarisation Thompson Sampling for multi-objective decisions in public media
This work provides a method for public service media to dynamically balance competing editorial objectives, improving the contextual relevance of their recommendations.
This paper addresses the challenge of multi-objective decision-making in public service media, where various competing objectives like audience reach and cultural values need balancing. The authors propose Contextual Scalarisation Thompson Sampler (CSTS), which learns to adaptively weight objectives based on context, demonstrating improved contextual relevance and alignment with expert practices on real programming data from Radio Télévision Suisse compared to existing methods.
Recommender systems may operate under multiple, competing objectives. For example, audience reach, cultural values, public service mandate, and operational constraints must be balanced in editorial decisions of public service media. Existing approaches relying on fixed combinations of objectives or Pareto-based optimisation do not adapt to changing priorities across situations. In this paper, we propose Contextual Scalarisation Thompson Sampler (CSTS), a multi-objective contextual bandit method that learns to weight objectives as a function of the observed context. We evaluate CSTS on real programming data from Radio Télévision Suisse, the Swiss national broadcaster, showing improved contextual relevance and better alignment with expert curation practices compared to fixed weight and standard contextual bandit approaches.