Weizmann Logo
ECCC
Electronic Colloquium on Computational Complexity

Under the auspices of the Computational Complexity Foundation (CCF)

Login | Register | Classic Style



REPORTS > DETAIL:

Revision(s):

Revision #1 to TR08-091 | 21st February 2012 02:43

On The Power of Membership Queries in Agnostic Learning

RSS-Feed




Revision #1
Authors: Vitaly Feldman
Accepted on: 21st February 2012 02:43
Downloads: 3000
Keywords: 


Abstract:

We study the properties of the agnostic learning framework of Haussler (1992)and Kearns, Schapire and Sellie (1992). In particular, we address the question: is there any situation in which membership queries are useful in agnostic learning?

Our results show that the answer is negative for distribution-independent agnostic learning and positive for agnostic learning with respect to a specific marginal distribution. Namely, we give a simple proof that any concept class learnable agnostically by a distribution-independent algorithm with access to membership queries is also learnable agnostically without membership queries. This resolves an open problem posed by Kearns et al. (1992). For agnostic learning with respect to the uniform distribution over $\{0,1\}^n$ we show a concept class that is learnable with membership queries but computationally hard to learn from random examples alone (assuming that one-way functions exist).



Changes to previous version:

Revision in response to reviewer comments


Paper:

TR08-091 | 10th September 2008 00:00

On The Power of Membership Queries in Agnostic Learning





TR08-091
Authors: Vitaly Feldman
Publication: 6th October 2008 14:47
Downloads: 3219
Keywords: 


Abstract:

We study the properties of the agnostic learning framework of Haussler (1992)and Kearns, Schapire and Sellie (1992). In particular, we address the question: is there any situation in which membership queries are useful in agnostic learning?

Our results show that the answer is negative for distribution-independent agnostic learning and positive for agnostic learning with respect to a specific marginal distribution. Namely, we give a simple proof that any concept class learnable agnostically by a distribution-independent algorithm with access to membership queries is also learnable agnostically without membership queries. This resolves an open problem posed by Kearns et al. (1992). For agnostic learning with respect to the uniform distribution over $\{0,1\}^n$ we show a concept class that is learnable with membership queries but computationally hard to learn from random examples alone (assuming that one-way functions exist).



ISSN 1433-8092 | Imprint