Axel Abels

This research project proposes an innovative approach to enhancing Collective Decision-Making (CDM) systems by synergistically combining adaptive social information exchange with advanced aggregation methods. Targeted at CDM settings like crowdsourced fact-checking, medical diagnostics, and content moderation, our novel approach uniquely balances the enhancement of individual input through carefully managed social interactions with the optimization of collective output through advanced aggregation.

We focus on aggregation algorithms that move beyond simple statistical heuristics by incorporating online learning techniques to track, valuate, and adapt to individual performance over time. These algorithms dynamically weight inputs based on expertise, context, and susceptibility to bias, allowing the system to resist noise and manipulation. The framework will be explicitly designed to reduce the impact of biases such as overconfidence, conformity, and authority bias, which often undermine group accuracy. Beyond traditional CDM with human groups, benefits of advancing bias-aware aggregation methods expand to artificial or hybrid human-artificial collectives.

Central to our endeavor is the integration of theoretical modeling, experimental research, and algorithmic development to navigate the complexities of social dynamics in decision-making processes. Our focus is on mitigating the biases inherent in social information exchange, while integrating multi-objective optimization to accommodate diverse stakeholder priorities, thereby creating a robust, bias-resistant, and effective CDM framework.