Please use this identifier to cite or link to this item:
https://hdl.handle.net/20.500.14279/27106
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Panousis, Konstantinos P. | - |
dc.contributor.author | Chatzis, Sotirios P. | - |
dc.contributor.author | Alexos, Antonios | - |
dc.contributor.author | Theodoridis, Sergios | - |
dc.date.accessioned | 2022-12-21T11:03:28Z | - |
dc.date.available | 2022-12-21T11:03:28Z | - |
dc.date.issued | 2021-01-04 | - |
dc.identifier.citation | International Conference on Artificial Intelligence and Statistics. PMLR, 2021. p. 3862-3870 | en_US |
dc.identifier.uri | https://hdl.handle.net/20.500.14279/27106 | - |
dc.description.abstract | This work addresses adversarial robustness in deep learning by considering deep networks with stochastic local winner-takes-all (LWTA) activations. This type of network units result in sparse representations from each model layer, as the units are organized in blocks where only one unit generates a non-zero output. The main operating principle of the introduced units lies on stochastic arguments, as the network performs posterior sampling over competing units to select the winner. We combine these LWTA arguments with tools from the field of Bayesian non-parametrics, specifically the stick-breaking construction of the Indian Buffet Process, to allow for inferring the sub-part of each layer that is essential for modeling the data at hand. Then, inference is performed by means of stochastic variational Bayes. We perform a thorough experimental evaluation of our model using benchmark datasets. As we show, our method achieves high robustness to adversarial perturbations, with state-of-the-art performance in powerful adversarial attack schemes. | en_US |
dc.language.iso | en | en_US |
dc.relation | aRTIFICIAL iNTELLIGENCE for the Deaf (aiD) | en_US |
dc.rights | Attribution-NonCommercial-NoDerivatives 4.0 International | * |
dc.rights.uri | http://creativecommons.org/licenses/by-nc-nd/4.0/ | * |
dc.subject | Computer Science - Learning | en_US |
dc.subject | Machine Learning | en_US |
dc.title | Local Competition and Stochasticity for Adversarial Robustness in Deep Learning | en_US |
dc.type | Conference Papers | en_US |
dc.collaboration | Cyprus University of Technology | en_US |
dc.collaboration | University of California | en_US |
dc.collaboration | National and Kapodistrian University of Athens | en_US |
dc.subject.category | Other Engineering and Technologies | en_US |
dc.journals | Open Access | en_US |
dc.country | Cyprus | en_US |
dc.country | United States | en_US |
dc.country | Greece | en_US |
dc.subject.field | Engineering and Technology | en_US |
dc.publication | Peer Reviewed | en_US |
dc.relation.conference | International Conference on Artificial Intelligence and Statistics | en_US |
dc.identifier.url | http://arxiv.org/abs/2101.01121v2 | - |
cut.common.academicyear | 2020-2021 | en_US |
item.grantfulltext | open | - |
item.languageiso639-1 | en | - |
item.openairecristype | http://purl.org/coar/resource_type/c_c94f | - |
item.openairetype | conferenceObject | - |
item.cerifentitytype | Publications | - |
item.fulltext | With Fulltext | - |
crisitem.project.funder | EC Joint Research Centre | - |
crisitem.project.fundingProgram | H2020 | - |
crisitem.project.openAire | info:eu-repo/grantAgreement/EC/H2020/872139 | - |
crisitem.author.dept | Department of Electrical Engineering, Computer Engineering and Informatics | - |
crisitem.author.faculty | Faculty of Engineering and Technology | - |
crisitem.author.orcid | 0000-0002-4956-4013 | - |
crisitem.author.parentorg | Faculty of Engineering and Technology | - |
Appears in Collections: | Δημοσιεύσεις σε συνέδρια /Conference papers or poster or presentation |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Local_competition.pdf | 612.4 kB | Adobe PDF | View/Open |
CORE Recommender
This item is licensed under a Creative Commons License