Improved Learning of AC0 Functions

Two extensions of the Linial, Mansour, Nisan AC0 learning algorithm are presented. The LMN method works when input examples are drawn uniformly. The new algorithmsimprove on theirs by performing well when given inputs drawn from unknown, mutually independent distributions. A variant of the one of the algorithms is conjectured to work in an even broader setting.

• Type: Conference
• Year: 1991
• Where: COLT
• Authors: Merrick L. Furst, Jeffrey C. Jackson, Sean W. Smith
• Keywords: COLT 1991, Input Examples, LMN Method, Machine Learning, Mutually Independent