Are generative deep models for novelty detection truly better?

Many deep models have been recently proposed for anomaly detection. This paper presents comparison of selected generative deep models and classical anomaly detection methods on an extensive number of non--image benchmark datasets. We provide statistical comparison of the selected models, in many con...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2018-07
Main Authors: Škvára, Vít, Pevný, Tomáš, Šmídl, Václav
Format: Article
Language:English
Subjects:
Anomalies
Image detection
Statistical methods
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites
container_end_page
container_issue
container_start_page
container_title arXiv.org
container_volume
creator Škvára, Vít
Pevný, Tomáš
Šmídl, Václav
description Many deep models have been recently proposed for anomaly detection. This paper presents comparison of selected generative deep models and classical anomaly detection methods on an extensive number of non--image benchmark datasets. We provide statistical comparison of the selected models, in many configurations, architectures and hyperparamaters. We arrive to conclusion that performance of the generative models is determined by the process of selection of their hyperparameters. Specifically, performance of the deep generative models deteriorates with decreasing amount of anomalous samples used in hyperparameter selection. In practical scenarios of anomaly detection, none of the deep generative models systematically outperforms the kNN.
format article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2073365707</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2073365707</sourcerecordid><originalsourceid>FETCH-proquest_journals_20733657073</originalsourceid><addsrcrecordid>eNqNikEKwjAQAIMgWLR_WPBciIltPSkiig_wXmq7lZaY1M2m0N-bgw_wNDAzC5EorXfZYa_USqTeD1JKVZQqz3UijmdCeKFFqrmfEFrEEd6uReOhcwTWTWh4jp6x4d5ZYApmhicyI502YtnVxmP641psb9fH5Z6N5D4BPVeDC2RjqpQstS7yMuK_6wvUaziT</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2073365707</pqid></control><display><type>article</type><title>Are generative deep models for novelty detection truly better?</title><source>Publicly Available Content Database</source><creator>Škvára, Vít ; Pevný, Tomáš ; Šmídl, Václav</creator><creatorcontrib>Škvára, Vít ; Pevný, Tomáš ; Šmídl, Václav</creatorcontrib><description>Many deep models have been recently proposed for anomaly detection. This paper presents comparison of selected generative deep models and classical anomaly detection methods on an extensive number of non--image benchmark datasets. We provide statistical comparison of the selected models, in many configurations, architectures and hyperparamaters. We arrive to conclusion that performance of the generative models is determined by the process of selection of their hyperparameters. Specifically, performance of the deep generative models deteriorates with decreasing amount of anomalous samples used in hyperparameter selection. In practical scenarios of anomaly detection, none of the deep generative models systematically outperforms the kNN.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Anomalies ; Image detection ; Statistical methods</subject><ispartof>arXiv.org, 2018-07</ispartof><rights>2018. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://www.proquest.com/docview/2073365707?pq-origsite=primo$$EHTML$$P50$$Gproquest$$Hfree_for_read</linktohtml><link.rule.ids>780,784,25753,37012,44590</link.rule.ids></links><search><creatorcontrib>Škvára, Vít</creatorcontrib><creatorcontrib>Pevný, Tomáš</creatorcontrib><creatorcontrib>Šmídl, Václav</creatorcontrib><title>Are generative deep models for novelty detection truly better?</title><title>arXiv.org</title><description>Many deep models have been recently proposed for anomaly detection. This paper presents comparison of selected generative deep models and classical anomaly detection methods on an extensive number of non--image benchmark datasets. We provide statistical comparison of the selected models, in many configurations, architectures and hyperparamaters. We arrive to conclusion that performance of the generative models is determined by the process of selection of their hyperparameters. Specifically, performance of the deep generative models deteriorates with decreasing amount of anomalous samples used in hyperparameter selection. In practical scenarios of anomaly detection, none of the deep generative models systematically outperforms the kNN.</description><subject>Anomalies</subject><subject>Image detection</subject><subject>Statistical methods</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2018</creationdate><recordtype>article</recordtype><sourceid>PIMPY</sourceid><recordid>eNqNikEKwjAQAIMgWLR_WPBciIltPSkiig_wXmq7lZaY1M2m0N-bgw_wNDAzC5EorXfZYa_USqTeD1JKVZQqz3UijmdCeKFFqrmfEFrEEd6uReOhcwTWTWh4jp6x4d5ZYApmhicyI502YtnVxmP641psb9fH5Z6N5D4BPVeDC2RjqpQstS7yMuK_6wvUaziT</recordid><startdate>20180713</startdate><enddate>20180713</enddate><creator>Škvára, Vít</creator><creator>Pevný, Tomáš</creator><creator>Šmídl, Václav</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20180713</creationdate><title>Are generative deep models for novelty detection truly better?</title><author>Škvára, Vít ; Pevný, Tomáš ; Šmídl, Václav</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_20733657073</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2018</creationdate><topic>Anomalies</topic><topic>Image detection</topic><topic>Statistical methods</topic><toplevel>online_resources</toplevel><creatorcontrib>Škvára, Vít</creatorcontrib><creatorcontrib>Pevný, Tomáš</creatorcontrib><creatorcontrib>Šmídl, Václav</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science &amp; Engineering Collection</collection><collection>ProQuest Central (Alumni)</collection><collection>ProQuest Central</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection (Proquest) (PQ_SDU_P3)</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Publicly Available Content Database</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Škvára, Vít</au><au>Pevný, Tomáš</au><au>Šmídl, Václav</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Are generative deep models for novelty detection truly better?</atitle><jtitle>arXiv.org</jtitle><date>2018-07-13</date><risdate>2018</risdate><eissn>2331-8422</eissn><abstract>Many deep models have been recently proposed for anomaly detection. This paper presents comparison of selected generative deep models and classical anomaly detection methods on an extensive number of non--image benchmark datasets. We provide statistical comparison of the selected models, in many configurations, architectures and hyperparamaters. We arrive to conclusion that performance of the generative models is determined by the process of selection of their hyperparameters. Specifically, performance of the deep generative models deteriorates with decreasing amount of anomalous samples used in hyperparameter selection. In practical scenarios of anomaly detection, none of the deep generative models systematically outperforms the kNN.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier EISSN: 2331-8422
ispartof arXiv.org, 2018-07
issn 2331-8422
language eng
recordid cdi_proquest_journals_2073365707
source Publicly Available Content Database
subjects Anomalies
Image detection
Statistical methods
title Are generative deep models for novelty detection truly better?
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-04T07%3A31%3A22IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=Are%20generative%20deep%20models%20for%20novelty%20detection%20truly%20better?&rft.jtitle=arXiv.org&rft.au=%C5%A0kv%C3%A1ra,%20V%C3%ADt&rft.date=2018-07-13&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2073365707%3C/proquest%3E%3Cgrp_id%3Ecdi_FETCH-proquest_journals_20733657073%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=2073365707&rft_id=info:pmid/&rfr_iscdi=true