On biased random walks, corrupted intervals, and learning under adversarial design

Daniel Berend; Aryeh Kontorovich; Lev Reyzin; Thomas Robinson

doi:10.1007/s10472-020-09696-1

On biased random walks, corrupted intervals, and learning under adversarial design

Daniel Berend, Aryeh Kontorovich, Lev Reyzin, Thomas Robinson

Department of Computer Science

Research output: Contribution to journal › Article › peer-review

Abstract

We tackle some fundamental problems in probability theory on corrupted random processes on the integer line. We analyze when a biased random walk is expected to reach its bottommost point and when intervals of integer points can be detected under a natural model of noise. We apply these results to problems in learning thresholds and intervals under a new model for learning under adversarial design.

Original language	English
Pages (from-to)	887-905
Number of pages	19
Journal	Annals of Mathematics and Artificial Intelligence
Volume	88
Issue number	8
DOIs	https://doi.org/10.1007/s10472-020-09696-1
State	Published - 1 Aug 2020

Keywords

Adversarial learning
Classification noise
Random walks

ASJC Scopus subject areas

Applied Mathematics
Artificial Intelligence

Access to Document

10.1007/s10472-020-09696-1

Cite this

@article{188b19ac143e47b9860e6efb48834355,

title = "On biased random walks, corrupted intervals, and learning under adversarial design",

abstract = "We tackle some fundamental problems in probability theory on corrupted random processes on the integer line. We analyze when a biased random walk is expected to reach its bottommost point and when intervals of integer points can be detected under a natural model of noise. We apply these results to problems in learning thresholds and intervals under a new model for learning under adversarial design.",

keywords = "Adversarial learning, Classification noise, Random walks",

author = "Daniel Berend and Aryeh Kontorovich and Lev Reyzin and Thomas Robinson",

note = "Publisher Copyright: {\textcopyright} 2020, Springer Nature Switzerland AG.",

year = "2020",

month = aug,

day = "1",

doi = "10.1007/s10472-020-09696-1",

language = "English",

volume = "88",

pages = "887--905",

journal = "Annals of Mathematics and Artificial Intelligence",

issn = "1012-2443",

publisher = "Springer Netherlands",

number = "8",

}

TY - JOUR

T1 - On biased random walks, corrupted intervals, and learning under adversarial design

AU - Berend, Daniel

AU - Kontorovich, Aryeh

AU - Reyzin, Lev

AU - Robinson, Thomas

PY - 2020/8/1

Y1 - 2020/8/1

N2 - We tackle some fundamental problems in probability theory on corrupted random processes on the integer line. We analyze when a biased random walk is expected to reach its bottommost point and when intervals of integer points can be detected under a natural model of noise. We apply these results to problems in learning thresholds and intervals under a new model for learning under adversarial design.

AB - We tackle some fundamental problems in probability theory on corrupted random processes on the integer line. We analyze when a biased random walk is expected to reach its bottommost point and when intervals of integer points can be detected under a natural model of noise. We apply these results to problems in learning thresholds and intervals under a new model for learning under adversarial design.

KW - Adversarial learning

KW - Classification noise

KW - Random walks

UR - http://www.scopus.com/inward/record.url?scp=85086025374&partnerID=8YFLogxK

U2 - 10.1007/s10472-020-09696-1

DO - 10.1007/s10472-020-09696-1

M3 - Article

AN - SCOPUS:85086025374

SN - 1012-2443

VL - 88

SP - 887

EP - 905

JO - Annals of Mathematics and Artificial Intelligence

JF - Annals of Mathematics and Artificial Intelligence

IS - 8

ER -

On biased random walks, corrupted intervals, and learning under adversarial design

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this