Pareto frontier

The Pareto frontier, named after the Italian economist Vilfredo Pareto, is the boundary formed by Pareto-optimal options. It is used in multi-objective optimization, where no option on the frontier is outperformed on every objective by another option.

In AI, the term is used to compare models across competing measures such as performance and cost or accuracy and latency. A model sits on the Pareto frontier when no other model delivers better performance at the same or lower cost.

This entry is based on AIPOST articles and widely known facts. If something is wrong, please send us a correction request.