Grounding Synthesis in Human Votes: Improving Sparse Matrix Completion for Opinion Selection in Large-Scale Consultations
Date:
Invited by the ParliView research team @ UC Dublin.
Getting insights from large scale citizen consultations is not trivial. Existing methods strongly rely on the Voting Matrix of a consultation, where citizens vote that they “agree” or “disagree” on each other’s propositions. However, this matrix is usually extremely sparse. algorithmically filling this matrix is a task known as Elicitation Inference Optimization (EIO). In this presentation, we explore different recommender system for this task.