Feature-based Uncertainty Model for School Choice

Researchers have proposed a new 'feature-based uncertainty' AI model for school choice. It helps students rank colleges when they can't form exact preferences, using features like reputation and location. The model treats a student's priorities as random variables. The proposed algorithm achieves a worst-case approximation ratio of (1/n)^n for stability and guarantees the strongest form of incentive compatibility, though these two goals are generally incompatible in matching theory.