Sparsity-Inducing Binarized Neural Networks
Open accessThe DOI matches its Crossref record; 16 of the 16 references it cites were found in a registry or an index.
Abstract
Binarization of feature representation is critical for Binarized Neural Networks (BNNs). Currently, sign function is the commonly used method for feature binarization. Although it works well on small datasets, the performance on ImageNet remains unsatisfied. Previous methods mainly focus on minimizing quantization error, improving the training strategies and decomposing each convolution layer into several binary convolution modules. However, whether sign is the only option for binarization has been largely overlooked. In this work, we propose the Sparsity-inducing Binarized Neural Network (Si-BNN), to quantize the activations to be either 0 or +1, which introduces sparsity into binary representation. We further introduce trainable thresholds into the backward function of binarization to guide the gradient propagation. Our method dramatically outperforms current state-of-the-arts, lowering the performance gap between full-precision networks and BNNs on mainstream architectures, achieving the new state-of-the-art on binarized AlexNet (Top-1 50.5%), ResNet-18 (Top-1 59.7%), and VGG-Net (Top-1 63.2%). At inference time, Si-BNN still enjoys the high efficiency of exclusive-not-or (xnor) operations.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333
XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks
Rastegari, M., Ordonez, V., Redmon, J., & Farhadi, A. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46493-0_32 · Cited by 4,408
BinaryConnect: Training Deep Neural Networks with binary weights during propagations
Courbariaux, M., Bengio, Y., & David, J.-P. (2015) · arXiv · doi:10.48550/arxiv.1511.00363 · Cited by 2,180
DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients
Zhou, S., Wu, Y., Ni, Z., Zhou, X., Wen, H., & Zou, Y. (2016) · arXiv · doi:10.48550/arxiv.1606.06160 · Cited by 1,690 · cited 1×
“Sinceweights and activations binarization leads to inference acceleration, DoReFa-Net (Zhou et al. 2016) further proposes the 2-bit gradients approximation with 1-bit weights and activations to boost training on FPGA.”
MethodHubara, I., Soudry, D., & Yaniv, R. E. (2016) · arXiv · doi:10.48550/arxiv.1602.02505 · Cited by 1,667
Deep Learning with Limited Numerical Precision
Gupta, S., Agrawal, A., Gopalakrishnan, K., & Narayanan, P. (2015) · arXiv · doi:10.48550/arxiv.1502.02551 · Cited by 1,364 · cited 1×
“It is shown that full-precision is not required in achieving high performance for deep neural networks(Gupta et al. 2015; Courbariaux, Bengio, and David 2015; Rastegari et al. 2016; Wang and Cheng 2017).”
BackgroundDeep Learning with Low Precision by Half-Wave Gaussian Quantization
Cai, Z., He, X., Sun, J., & Vasconcelos, N. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.574 · Cited by 498 · cited 13×
“This finding enable us to further explore sparse feature representation, which have shown to be effective (Cai et al. 2017; Jung et al. 2019).”
BackgroundMethodInfluentialSelf-citation
Cited by
Improving Extreme Low-Bit Quantization With Soft Threshold
Xu, W., Li, F., Jiang, Y., Yong, A., He, X., Wang, P., & Cheng, J. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2022.3216389 · Cited by 32
“As for ImageNet experiments, following [50], [52], [53], we first proportionally resize images to 256 × N with the short edge to 256.”
MethodSelf-citationA Novel Mixed-Precision Quantization Approach for CNNs
Wu, D., Wang, Y., Fei, Y., & Gao, G. (2025) · IEEE Access · doi:10.1109/access.2025.3551802 · Cited by 8
“A number of subsequent quantization methods such as Wang et al. [27] and Ding et al. [28] have presented results showing improving upon this model to varying degrees.”
Oflamaz, C. U., & Yalçın, M. E. (2025) · Electronics · doi:10.3390/electronics14183686 · Cited by 4
Domain-Adaptive Online Active Learning for Real-Time Intelligent Video Analytics on Edge Devices
Boldo, M., De Marchi, M., Martini, E., Aldegheri, S., & Bombieri, N. (2024) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2024.3453188 · Cited by 4
“Examples are model compression techniques, such as neural network pruning [5], quantization [6], compact network design [7], as well as strategies like early exit [8].”
MethodSTBLLM: Breaking the 1-Bit Barrier with Structured Binary LLMs
Dong, P., Li, L., Zhong, Y., Du, D., Fan, R., Chen, Y., Tang, Z., Wang, Q., Xue, W., Guo, Y., et al. (2024) · arXiv · doi:10.48550/arxiv.2408.01803 · Cited by 1
Metadata
- Authors
- Wang, P., He, X., Li, G., Zhao, T., & Cheng, J.
- Venue
- Proceedings of the AAAI Conference on Artificial Intelligence
- Published
- 2020-04-03
- Volume
- 34
- Issue
- 07
- Pages
- 12192-12199
- ISSN
- 2159-5399 (ISSN-L)
- 2374-3468
- OpenAlex
- W2998419362
- Cited by
- 49
- References
- 16
- FWCI
- 2.60
- Type
- conference paper
- Language
- en
- Open access
- diamond
- Licence
- www.aaai.org/ (unspecified)
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.