• implement the regular positive and negative inference (RPNI) algorithm of Oncina and Garcia, as described here and empirically expore its behavior.
  • learning random shuffle ideals — either theoretically or empirically