Zur Kurzanzeige

dc.contributor.author
Fischer, Marc
dc.contributor.author
Mirman, Matthew
dc.contributor.author
Stalder, Steven
dc.contributor.author
Vechev, Martin
dc.date.accessioned
2020-05-20T11:37:46Z
dc.date.available
2020-01-29T12:06:15Z
dc.date.available
2020-05-20T11:37:46Z
dc.date.issued
2019-11-22
dc.identifier.uri
http://hdl.handle.net/20.500.11850/395397
dc.description.abstract
In deep reinforcement learning (RL), adversarial attacks can trick an agent into unwanted states and disrupt training. We propose a system called Robust StudentDQN (RS-DQN), which permits online robustness training alongside Q networks, while preserving competitive performance. We show that RS-DQN can be combined with (i) state-of-the-art adversarial training and (ii) provably robust training to obtain an agent that is resilient to strong attacks during training and evaluation.
en_US
dc.language.iso
en
en_US
dc.publisher
Cornell University
en_US
dc.title
Online Robustness Training for Deep Reinforcement Learning
en_US
dc.type
Working Paper
ethz.journal.title
arXiv
ethz.pages.start
1911.00887v3
en_US
ethz.size
14 p.
en_US
ethz.event
arXiv:1903.12519v2 [
en_US
ethz.identifier.arxiv
1911.00887v3
ethz.publication.place
Ithaca, NY
en_US
ethz.publication.status
published
en_US
ethz.leitzahl
ETH Zürich::00002 - ETH Zürich::00012 - Lehre und Forschung::00007 - Departemente::02150 - Dep. Informatik / Dep. of Computer Science::02664 - Inst. f. Programmiersprachen u. -systeme / Inst. Programming Languages and Systems::03948 - Vechev, Martin / Vechev, Martin
en_US
ethz.leitzahl.certified
ETH Zürich::00002 - ETH Zürich::00012 - Lehre und Forschung::00007 - Departemente::02150 - Dep. Informatik / Dep. of Computer Science::02664 - Inst. f. Programmiersprachen u. -systeme / Inst. Programming Languages and Systems::03948 - Vechev, Martin / Vechev, Martin
en_US
ethz.date.deposited
2020-01-29T12:06:22Z
ethz.source
FORM
ethz.eth
yes
en_US
ethz.availability
Metadata only
en_US
ethz.rosetta.installDate
2020-05-20T11:37:55Z
ethz.rosetta.lastUpdated
2020-05-20T11:37:55Z
ethz.rosetta.versionExported
true
ethz.COinS
ctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.atitle=Online%20Robustness%20Training%20for%20Deep%20Reinforcement%20Learning&rft.jtitle=arXiv&rft.date=2019-11-22&rft.spage=1911.00887v3&rft.au=Fischer,%20Marc&Mirman,%20Matthew&Stalder,%20Steven&Vechev,%20Martin&rft.genre=preprint&
 Printexemplar via ETH-Bibliothek suchen

Dateien zu diesem Eintrag

DateienGrößeFormatIm Viewer öffnen

Zu diesem Eintrag gibt es keine Dateien.

Publikationstyp

Zur Kurzanzeige