Exact learning from membership queries: Some techniques, results and new directions

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Given a black box that contains a function f:D → R from some class of functions C. The black box can receive an element d (query) of the domain D and in time T returns the value f(d) ∈ R. Our goal is to exactly find (exactly learn) f with minimum number of queries and optimal time complexity. Or at least decide whether f ≡ g for some function g ∈ C. This problem has different names in different areas: Interpolation, Exactly Learning, Inferring, Identifying, Active Learning, Guessing Game, Testing, Functional Verification, Hitting Set and Black Box PIT from Substitution or Membership Queries. In this survey we give some of the results known from the literature, different techniques used mainly for the problem of exact learning and new directions that we think are worth investigating.

Original languageEnglish
Title of host publicationAlgorithmic Learning Theory - 24th International Conference, ALT 2013, Proceedings
Pages33-52
Number of pages20
DOIs
StatePublished - 2013
Event24th International Conference on Algorithmic Learning Theory, ALT 2013 - Singapore, Singapore
Duration: 6 Oct 20139 Oct 2013

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8139 LNAI

Conference

Conference24th International Conference on Algorithmic Learning Theory, ALT 2013
Country/TerritorySingapore
CitySingapore
Period6/10/139/10/13

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'Exact learning from membership queries: Some techniques, results and new directions'. Together they form a unique fingerprint.

Cite this