Filters








20,043 Hits in 4.7 sec

Learning with errors in answers to membership queries

Laurence Bisht, Nader H. Bshouty, Lawrance Khoury
2008 Journal of computer and system sciences (Print)  
Turán, Malicious omissions and errors in answering to membership queries, Machine Learning 28 (2-3) (1997) 211-255]: Learning with equivalence and limited membership queries and learning with equivalence  ...  This closes the open problems in [D. Angluin, M. Krikis, R.H. Sloan, G. Turán, Malicious omissions and errors in answering to membership queries, Machine Learning 28 (2-3) (1997) 211-255].  ...  In this model the learning algorithm learns exactly a target function using equivalence and membership queries with at most some number of errors or omissions in answers to the membership queries.  ... 
doi:10.1016/j.jcss.2007.04.010 fatcat:zv5sqa3adreexdznasdyeoztam

Randomly fallible teachers: Learning monotone DNF with an incomplete membership oracle

Dana Angluin, Donna K. Slonim
1994 Machine Learning  
We introduce a new fault-tolerant model of algorithmic learning using an equivalence oracle and an incomplete membership oracle, in which the answers to a random subset of the learner's membership queries  ...  We demonstrate that, with high probability, it is still possible to learn monotone DNF formulas in polynomial time, provided that the fraction of missing answers is bounded by some constant less than one  ...  A preliminary version of this paper appeared in the proceedings of the 1991 Workshop on Computational Learning Theory (Angluin and Slonim, 1991) .  ... 
doi:10.1007/bf00993160 fatcat:clgi4bcec5eivmrqhlsuuw63n4

Learning the Language of Software Errors

Hana Chockler, Pascal Kesseli, Daniel Kroening, Ofer Strichman
2020 The Journal of Artificial Intelligence Research  
We propose to use algorithms for learning deterministic finite automata (DFA), such as Angluin's L* algorithm, for learning a DFA that describes the possible scenarios under which a given program error  ...  We also present lazy learning, which is a method for reducing the number of membership queries while using L*, and demonstrate its effectiveness on standard benchmarks.  ...  the answers to the membership queries answered so far.  ... 
doi:10.1613/jair.1.11798 fatcat:wc7xx2ihcbanljxzs2an32nrve

Learning the Language of Error [chapter]

Martin Chapman, Hana Chockler, Pascal Kesseli, Daniel Kroening, Ofer Strichman, Michael Tautschnig
2015 Lecture Notes in Computer Science  
We propose to harness Angluin's L * algorithm for learning a deterministic finite automaton that describes the possible scenarios under which a given program error occurs.  ...  This can be used, for example, for visually comparing different versions of a program, by presenting an automaton for the behaviour in the symmetric difference between them, or for assisting in merging  ...  We bypass a CBMC call and answer 'false' to a membership query if one of the following holds: -The query does not end with a call to assert, -The query contains more than one call to assert, w is incompatible  ... 
doi:10.1007/978-3-319-24953-7_9 fatcat:3kvt5kst35fnlhcn2watfnic3i

Separating models of learning with faulty teachers

Vitaly Feldman, Shrenik Shah
2009 Theoretical Computer Science  
In this model, the answers to a random subset of the learner's membership queries are flipped.  ...  In this model, the answers to a random subset of the learner's membership queries may be missing.  ...  We are grateful to Salil Vadhan for contributing an important insight into the proof of Theorem 1.1. We also thank Neal Wadhwa for his help in the write-up of an early version of these results.  ... 
doi:10.1016/j.tcs.2009.01.017 fatcat:mojru55gsjeoxdkxh5z22tnuk4

Separating Models of Learning with Faulty Teachers [chapter]

Vitaly Feldman, Shrenik Shah, Neal Wadhwa
2007 Lecture Notes in Computer Science  
In this model, the answers to a random subset of the learner's membership queries are flipped.  ...  In this model, the answers to a random subset of the learner's membership queries may be missing.  ...  We are grateful to Salil Vadhan for contributing an important insight to the proof of Theorem 1.1. We also thank Neal Wadhwa for his help in the write-up of an early version of these results.  ... 
doi:10.1007/978-3-540-75225-7_11 fatcat:ak4o5c6cevd5riazvzzr5p2eiq

A general dimension for query learning

José L. Balcázar, Jorge Castro, David Guijarro, Johannes Köbler, Wolfgang Lindner
2007 Journal of computer and system sciences (Print)  
Further, in the approximate learning setting, we use the general dimension to characterize the query complexity in the statistical query as well as the learning by distances model.  ...  We introduce a combinatorial dimension that characterizes the number of queries needed to exactly (or approximately) learn concept classes in various models.  ...  Shai Ben-David discussed with one of the authors the learning by distance model and its relationship with the statistical query model. Also, we thank the referees for their useful comments.  ... 
doi:10.1016/j.jcss.2007.03.003 fatcat:uz5ths4rw5f6pmmjxiaobfg3d4

How Many Query Superpositions Are Needed to Learn? [chapter]

Jorge Castro
2006 Lecture Notes in Computer Science  
This paper introduces a framework for quantum exact learning via queries, the so-called quantum protocol. It is shown that usual protocols in the classical learning setting have quantum counterparts.  ...  Given a quantum protocol and a target concept class, the general halving dimension provides lower and upper bounds on the number of queries that a quantum algorithm needs to learn.  ...  Acknowledgments This work was supported in part by the IST Programme of the European Community, under the PASCAL Network of Excellence, IST-2002-506778, and by the spanish MCYT research project TRANGRAM  ... 
doi:10.1007/11894841_10 fatcat:v2v7ykwxjzenhhvgvn3cserk3e

Learning First-Order Definite Theories via Object-Based Queries [chapter]

Joseph Selman, Alan Fern
2011 Lecture Notes in Computer Science  
These results show, in theory, the potential for incorporating object-based queries into first-order learning algorithms in order to reduce human teaching effort.  ...  Prior work has shown that first order Horn theories can be learned using a polynomial number of membership and equivalence queries [6] .  ...  Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the DARPA.  ... 
doi:10.1007/978-3-642-23808-6_11 fatcat:qzl4ivto55hwrkawaddjqrtlc4

Queries and concept learning

Dana Angluin
1988 Machine Learning  
We consider the problem of using queries to learn an unknown concept.  ...  Several types of queries are described and studied: membership, equivalence, subset, superset, disjointness, and exhaustiveness queries.  ...  This paper is a revision of an earlier report (Angluin, 1986 ), but is not identical to it.  ... 
doi:10.1007/bf00116828 fatcat:5f2oypi2abbubkcatlw6hqenyu

Active Learning with Generalized Queries

Jun Du, Charles X. Ling
2009 2009 Ninth IEEE International Conference on Data Mining  
In this paper, we propose a novel active learning algorithm that asks generalized queries.  ...  A more natural way is to ask "generalized queries" with don't-care attributes, such as "are people over 50 with knee pain likely to have osteoarthritis?"  ...  ACKNOWLEDGMENT The authors acknowledge the valuable assistance of other members of the Data Mining and E-Business Lab of The University of Western Ontario in this research.  ... 
doi:10.1109/icdm.2009.71 dblp:conf/icdm/DuL09 fatcat:n2hnenjmaremlh26zfwts6hjle

DataPlay

Azza Abouzied, Joseph Hellerstein, Avi Silberschatz
2012 Proceedings of the 25th annual ACM symposium on User interface software and technology - UIST '12  
Users receive real-time feedback in the form of updated answers and non-answers.  ...  We identify two impediments to this form of interactive trial-and-error query specification in SQL: (i) changing quantifiers often requires global syntactical query restructuring, and (ii) the absence  ...  Enabling users to label offending tuples in answers with 'want out' and actual answers misplaced in non-answers with 'want in' and in turn correcting the query for the user reduces the gulf of execution  ... 
doi:10.1145/2380116.2380144 dblp:conf/uist/AbouziedHS12 fatcat:vehrq7irvbhb7jfihazzpnuxxy

:{unav)

Dana Angluin
2012 Machine Learning  
We consider the problem of using queries to learn an unknown concept.  ...  Several types of queries are described and studied: membership, equivalence, subset, superset, disjointness, and exhaustiveness queries.  ...  This paper is a revision of an earlier report (Angluin, 1986 ), but is not identical to it.  ... 
doi:10.1023/a:1022821128753 fatcat:q4euxoaxsza65bhf4fk2gudbbi

PAC Learning-Based Verification and Model Synthesis [article]

Yu-Fang Chen, Chiao Hsieh, Ondřej Lengál, Tsung-Ju Lii, Ming-Hsien Tsai, Bow-Yaw Wang, Farn Wang
2015 arXiv   pre-print
Our learning procedure is therefore based on the framework of probably approximately correct (PAC) learning, which uses sampling instead and provides correctness guarantees expressed using the terms error  ...  Exact learning algorithms require checking equivalence between the model and the program, which is a difficult problem, in general undecidable.  ...  As with other online learning-based techniques, we need to devise a mechanical teacher that answers queries from the learning algorithm.  ... 
arXiv:1511.00754v1 fatcat:spsjmw6xa5cf7chgnherfodb4i

Learning orthogonal F-Horn formulas

Eiji Takimoto, Akira Miyashiro, Akira Maruoka, Yoshifumi Sakai
1997 Theoretical Computer Science  
Recently, it was pointed out that the problem of PAC-learning for these classes with membership queries can be reduced to that of learning for the class of k-quasi Horn formulas with membership and equivalence  ...  In the PAC-learning, or the query learning model, it has been an important open problem to decide whether the class of DNF and CNF formulas is learnable.  ...  answer of the query to the algorithm.  ... 
doi:10.1016/s0304-3975(97)00020-0 fatcat:o6hmxyj5rbhqlm6csuxbvvpj3a
« Previous Showing results 1 — 15 out of 20,043 results