Augmented Decision Making

Inspiration

Social choice has received considerable attention in AI and computer science in recent years, due to technological advances that have facilitated an explosion in the availability of ranking or preference data. Despite a great degree of personalization in recommendations and information provision, tailoring alternatives presented to specific users can be difficult for a number of reasons, including infeasibility of complete personalization. Most decisions force the choice of a single option , usually from a range of choices that maximise customer satisfaction across a target market. In such settings, a single “consensus” recommendation is made for the population as a whole.

In social choice theory, the Arrow's theorem states that when voters have three or more distinct alternatives , no ranked voting electoral system can convert the ranked preferences of individuals into a community-wide ranking while also meeting a specified set of criteria: unrestricted domain, non-dictatorship, Pareto efficiency, and independence of irrelevant alternatives

The proposed alternative is using a modern voting system so as to satisfy as many criteria as possible, while considering the participatory democracy concept: involving users in the decision making process and focusing on their participation and collaboration

What it does

Our platform offers users a middle ground between fully personalised choices and pure consensus recommendations in the process of decision making. The platform is an integrated part of the Mirro product and aims to benefit organisations by empowering individual employees in the fair-decision making process implied by a participatory democracy:

problem identification
debate , ideation and co-creation
drafting proposals
voting
checking that the solution has been properly deployed and actually solves the identified problem.

The process aims to determine a collection of candidates that best represent the “collective interests” of the voters, by using a modern voting system, selectable by problem authors when defining the problem.

By focusing on educated user collaboration in the context of ideation and debating, the platform offers a great opportunity for gathering learning data for a sentiment analysis ML model. Users should be able to mark their content as positive during the debate and ideation process. Even though datasets for sentiment analysis do exist, none of them are in Romanian.