ALVIN: Active Learning Via INterpolation