Balanced Binary Neural Networks with Gated Residual
Binary neural networks have attracted numerous attention in recent years. However, mainly due to the information loss stemming from the biased binarization, how to preserve the accuracy of networks still remains a critical issue. In this paper, we attempt to maintain the information propagated in th...
Saved in:
Main Authors | , , , |
---|---|
Format | Journal Article |
Language | English |
Published |
26.09.2019
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Abstract | Binary neural networks have attracted numerous attention in recent years.
However, mainly due to the information loss stemming from the biased
binarization, how to preserve the accuracy of networks still remains a critical
issue. In this paper, we attempt to maintain the information propagated in the
forward process and propose a Balanced Binary Neural Networks with Gated
Residual (BBG for short). First, a weight balanced binarization is introduced
to maximize information entropy of binary weights, and thus the informative
binary weights can capture more information contained in the activations.
Second, for binary activations, a gated residual is further appended to
compensate their information loss during the forward process, with a slight
overhead. Both techniques can be wrapped as a generic network module that
supports various network architectures for different tasks including
classification and detection. We evaluate our BBG on image classification tasks
over CIFAR-10/100 and ImageNet and on detection task over Pascal VOC. The
experimental results show that BBG-Net performs remarkably well across various
network architectures such as VGG, ResNet and SSD with the superior performance
over state-of-the-art methods in terms of memory consumption, inference speed
and accuracy. |
---|---|
AbstractList | Binary neural networks have attracted numerous attention in recent years.
However, mainly due to the information loss stemming from the biased
binarization, how to preserve the accuracy of networks still remains a critical
issue. In this paper, we attempt to maintain the information propagated in the
forward process and propose a Balanced Binary Neural Networks with Gated
Residual (BBG for short). First, a weight balanced binarization is introduced
to maximize information entropy of binary weights, and thus the informative
binary weights can capture more information contained in the activations.
Second, for binary activations, a gated residual is further appended to
compensate their information loss during the forward process, with a slight
overhead. Both techniques can be wrapped as a generic network module that
supports various network architectures for different tasks including
classification and detection. We evaluate our BBG on image classification tasks
over CIFAR-10/100 and ImageNet and on detection task over Pascal VOC. The
experimental results show that BBG-Net performs remarkably well across various
network architectures such as VGG, ResNet and SSD with the superior performance
over state-of-the-art methods in terms of memory consumption, inference speed
and accuracy. |
Author | Shen, Mingzhu Han, Kai Gong, Ruihao Liu, Xianglong |
Author_xml | – sequence: 1 givenname: Mingzhu surname: Shen fullname: Shen, Mingzhu – sequence: 2 givenname: Xianglong surname: Liu fullname: Liu, Xianglong – sequence: 3 givenname: Ruihao surname: Gong fullname: Gong, Ruihao – sequence: 4 givenname: Kai surname: Han fullname: Han, Kai |
BackLink | https://doi.org/10.48550/arXiv.1909.12117$$DView paper in arXiv |
BookMark | eNotzsFuwjAQBFAf2gNQPqAn8gMJXmzj-NhELUVCRaq4Rxt7LSzSUDmhlL8n0J5GI41Gb8we2mNLjD0Dz2SuFJ9j_A0_GRhuMlgA6BETBTbYWnJJEVqMl-SDThGbIfrzMR665Bz6fbLCflh8UhfcCZsn9uix6Wj6nxO2e3vdle_pZrtaly-bFJdap9I7pSkn4RfIJQEQWEIy3NYo1FBsrUhrghqsdCZ3pJ03SkmSFkyuxITN_m7v6uo7hq8BWN301V0vrqnyQXE |
ContentType | Journal Article |
Copyright | http://arxiv.org/licenses/nonexclusive-distrib/1.0 |
Copyright_xml | – notice: http://arxiv.org/licenses/nonexclusive-distrib/1.0 |
DBID | AKY GOX |
DOI | 10.48550/arxiv.1909.12117 |
DatabaseName | arXiv Computer Science arXiv.org |
DatabaseTitleList | |
Database_xml | – sequence: 1 dbid: GOX name: arXiv.org url: http://arxiv.org/find sourceTypes: Open Access Repository |
DeliveryMethod | fulltext_linktorsrc |
ExternalDocumentID | 1909_12117 |
GroupedDBID | AKY GOX |
ID | FETCH-LOGICAL-a677-4fd57e8e3f2a04e11e1ceae90cba35e1ccb5e77e1b1c4d98de7df9554e4c19853 |
IEDL.DBID | GOX |
IngestDate | Mon Jan 08 05:50:24 EST 2024 |
IsDoiOpenAccess | true |
IsOpenAccess | true |
IsPeerReviewed | false |
IsScholarly | false |
Language | English |
LinkModel | DirectLink |
MergedId | FETCHMERGED-LOGICAL-a677-4fd57e8e3f2a04e11e1ceae90cba35e1ccb5e77e1b1c4d98de7df9554e4c19853 |
OpenAccessLink | https://arxiv.org/abs/1909.12117 |
ParticipantIDs | arxiv_primary_1909_12117 |
PublicationCentury | 2000 |
PublicationDate | 2019-09-26 |
PublicationDateYYYYMMDD | 2019-09-26 |
PublicationDate_xml | – month: 09 year: 2019 text: 2019-09-26 day: 26 |
PublicationDecade | 2010 |
PublicationYear | 2019 |
Score | 1.750981 |
SecondaryResourceType | preprint |
Snippet | Binary neural networks have attracted numerous attention in recent years.
However, mainly due to the information loss stemming from the biased
binarization,... |
SourceID | arxiv |
SourceType | Open Access Repository |
SubjectTerms | Computer Science - Computer Vision and Pattern Recognition |
Title | Balanced Binary Neural Networks with Gated Residual |
URI | https://arxiv.org/abs/1909.12117 |
hasFullText | 1 |
inHoldings | 1 |
isFullTextHit | |
isPrint | |
link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwdV07T8MwED61nVgQCFB5ygNrRPyIHY8UUSqGIqEiZYv8uEgsCDUF8fM5O0GwMFm2b7nz8N3Z330GuPbCB-S-LGpVyyL1MhZOW19YjMrpTkYaEttirVcv6rGpmgmwn14Yt_16_Rz0gX1_Q2hlk_4BN1OYCpEoWw9PzfA4maW4RvtfO8ox89IfkFgewP6Y3bHb4TgOYYJvRyAXiT4YMLJFbn5lSRCDrNYDA7tn6S6UpWusyJ6xz91Rx7BZ3m_uVsX4VwG5ZkyhulgZrFF2wpUKOUce0KEtg3eyoknwFRpDUeFBRVtHNLGzBOWoApX9lTyBGZX7OAemlXFaUtCoMlEobC1siDp9KhVsbQw_hXn2sH0f5Cja5HybnT_7f-sc9gjqMztK6AuY7bYfeElwuvNXOabf4Xl0gQ |
link.rule.ids | 228,230,786,891 |
linkProvider | Cornell University |
openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Balanced+Binary+Neural+Networks+with+Gated+Residual&rft.au=Shen%2C+Mingzhu&rft.au=Liu%2C+Xianglong&rft.au=Gong%2C+Ruihao&rft.au=Han%2C+Kai&rft.date=2019-09-26&rft_id=info:doi/10.48550%2Farxiv.1909.12117&rft.externalDocID=1909_12117 |