In computational learning theory, Occam learning is a model of algorithmic learning where the objective of the learner is to output a succinct representation of received training data.
This is closely related to probably approximately correct (PAC) learning, where the learner is evaluated on its predictive power of a test set.
Occam learnability implies PAC learning, and for a wide variety of concept classes, the converse is also true: PAC learnability implies Occam learnability.
Occam Learning is named after Occam's razor, which is a principle stating that, given all other things being equal, a shorter explanation for observed data should be favored over a lengthier explanation.
The theory of Occam learning is a formal and mathematical justification for this principle.
In other words, parsimony (of the output hypothesis) implies predictive power.
of the shortest bit string that can represent
Occam learning connects the succinctness of a learning algorithm's output to its predictive power on unseen data.
is the maximum length of any sample
An Occam algorithm is called efficient if it runs in time polynomial in
is Occam learnable with respect to a hypothesis class
if there exists an efficient Occam algorithm for
Occam learnability implies PAC learnability, as the following theorem of Blumer, et al.[2] shows: Let
is also a PAC learner for the concept class
A slightly more general formulation is as follows: Let
samples drawn from a fixed but unknown distribution
.While the above theorems show that Occam learning is sufficient for PAC learning, it doesn't say anything about necessity.
Board and Pitt show that, for a wide variety of concept classes, Occam learning is in fact necessary for PAC learning.
[3] They proved that for any concept class that is polynomially closed under exception lists, PAC learnability implies the existence of an Occam algorithm for that concept class.
Concept classes that are polynomially closed under exception lists include Boolean formulas, circuits, deterministic finite automata, decision-lists, decision-trees, and other geometrically-defined concept classes.
is polynomially closed under exception lists if there exists a polynomial-time algorithm
of exceptions, outputs a representation of a concept
The probability that a set of samples
By the union bound, the probability that there exists a bad hypothesis in
This concludes the proof of the second theorem above.
-Occam algorithm, this means that any hypothesis output by
Thus, by the Cardinality version Theorem,
This concludes the proof of the first theorem above.
Though Occam and PAC learnability are equivalent, the Occam framework can be used to produce tighter bounds on the sample complexity of classical problems including conjunctions,[2] conjunctions with few relevant variables,[4] and decision lists.
[5] Occam algorithms have also been shown to be successful for PAC learning in the presence of errors,[6][7] probabilistic concepts,[8] function learning[9] and Markovian non-independent examples.