The DOI matches its Crossref record; 22 of the 22 references it cites were found in a registry or an index.
Comparing Kullback-Leibler Divergence and Mean Squared Error Loss in Knowledge Distillation
Open accessAbstract
Knowledge distillation (KD), transferring knowledge from a cumbersome teacher model to a lightweight student model, has been investigated to design efficient neural architectures. Generally, the objective function of KD is the Kullback-Leibler (KL) divergence loss between the softened probability distributions of the teacher model and the student model with the temperature scaling hyperparameter τ. Despite its widespread use, few studies have discussed how such softening influences generalization. Here, we theoretically show that the KL divergence loss focuses on the logit matching when τ increases and the label matching when τ goes to 0 and empirically show that the logit matching is positively correlated to performance improvement in general. From this observation, we consider an intuitive KD loss function, the mean squared error (MSE) between the logit vectors, so that the student model can directly learn the logit of the teacher model. The MSE loss outperforms the KL divergence loss, explained by the penultimate layer representations difference between the two losses. Furthermore, we show that sequential distillation can improve performance and that KD, using the KL divergence loss with small τ particularly, mitigates the label noise. The code to reproduce the experiments is publicly available online at https://github.com/jhoon-oh/kd_data/.
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
Rethinking the Inception Architecture for Computer Vision
Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.308 · Cited by 32,138
Distilling the Knowledge in a Neural Network
Hinton, G., Vinyals, O., & Dean, J. (2015) · arXiv · doi:10.48550/arxiv.1503.02531 · Cited by 14,204
Zagoruyko, S., & Komodakis, N. (2016) · Procedings of the British Machine Vision Conference 2016 · doi:10.5244/c.30.87 · Cited by 7,244
Categorical Reparameterization with Gumbel-Softmax
Jang, E., Gu, S., & Poole, B. (2016) · arXiv · doi:10.48550/arxiv.1611.01144 · Cited by 3,616
FitNets: Hints for Thin Deep Nets
Romero, A., Ballas, N., Kahou, S. E., Chassang, A., Gatta, C., & Bengio, Y. (2014) · arXiv · doi:10.48550/arxiv.1412.6550 · Cited by 2,859
Zagoruyko, S., & Komodakis, N. (2016) · arXiv · doi:10.48550/arxiv.1612.03928 · Cited by 2,027
Relational Knowledge Distillation
Park, W., Kim, D., Lu, Y., & Cho, M. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00409 · Cited by 1,681
The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables
Maddison, C. J., Mnih, A., & Teh, Y. W. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1611.00712 · Cited by 1,359
Do Deep Nets Really Need to be Deep?
Ba, J. L., & Caruana, R. (2013) · arXiv (Cornell University) · doi:10.48550/arxiv.1312.6184 · Cited by 1,335
Cited by
Alzubaidi, L., Bai, J., Al-Sabaawi, A., Santamaría, J., Albahri, A. S., Al-dabbagh, B. S. N., Fadhel, M. A., Manoufali, M., Zhang, J., Al-Timemy, A. H., et al. (2023) · Journal of Big Data · doi:10.1186/s40537-023-00727-2 · Cited by 907
Metadata
- Authors
- Kim, T., Oh, J., Kim, N. Y., Cho, S., & Yun, S.-Y.
- Venue
- Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence
- Published
- 2021-08
- Pages
- 2628-2635
- OpenAlex
- W3187295906
- Cited by
- 270
- References
- 22
- FWCI
- 13.25
- Type
- conference paper
- Language
- en
- Open access
- gold
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.