Abstract
Most state-of-the-art deep neural networks are overparameterized and exhibit a high computational cost. A straightforward approach to this problem is to replace convolutional kernels with its low-rank tensor approximations, whereas the Canonical Polyadic tensor Decomposition is one of the most suited models. However, fitting the convolutional tensors by numerical optimization algorithms often encounters diverging components, i.e., extremely large rank-one tensors but canceling each other. Such degeneracy often causes the non-interpretable result and numerical instability for the neural network ne-tuning. This paper is the first study on degeneracy in the tensor decomposition of convolutional kernels. We present a novel method, which can stabilize the low-rank approximation of convolutional kernels and ensure efficient compression while preserving the high-quality performance of the neural networks. We evaluate our approach on popular CNN architectures for image classification and show that our method results in much lower accuracy degradation and provides consistent performance.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
Notes
- 1.
Rank-1 tensor of size \(n_1\times n_2\times \dots \times n_{d}\) is an outer product of d vectors with dimensions \(n_1, n_1,\dots , n_d\).
- 2.
The mode-j unfolding of an order-d tensor of size \(n_1\times n_2 \times \dots \times n_d\) reorders the elements of the tensor into a matrix with \(n_j\) rows and \(n_1\dots n_{j - 1}n_{j + 1}\dots n_d\) columns.
- 3.
As shown in [53], RMS error is not the only one minimization criterion for a particular computer vision task.
References
Astrid, M., Lee, S.: CP-decomposition with tensor power method for convolutional neural networks compression. In: 2017 IEEE International Conference on Big Data and Smart Computing, BigComp 2017, Jeju Island, South Korea, 13–16 February 2017, pp. 115–118. IEEE (2017). https://doi.org/10.1109/BIGCOMP.2017.7881725
Bulat, A., Kossaifi, J., Tzimiropoulos, G., Pantic, M.: Matrix and tensor decompositions for training binary neural networks. arXiv preprint arXiv:1904.07852 (2019)
Bulat, A., Kossaifi, J., Tzimiropoulos, G., Pantic, M.: Incremental multi-domain learning with network latent tensor factorization. In: AAAI (2020)
Chen, T., Lin, J., Lin, T., Han, S., Wang, C., Zhou, D.: Adaptive mixture of low-rank factorizations for compact neural modeling. In: CDNNRIA Workshop, NIPS (2018)
Cichocki, A., Lee, N., Oseledets, I., Phan, A.H., Zhao, Q., Mandic, D.P.: Tensor networks for dimensionality reduction and large-scale optimization: Part 1 low-rank tensor decompositions. Found. Trends\(\textregistered \) Mach. Learn. 9(4–5), 249–429 (2016)
De Lathauwer, L.: Decompositions of a higher-order tensor in block terms – Part I and II. SIAM J. Matrix Anal. Appl. 30(3), 1022–1066 (2008). http://publi-etis.ensea.fr/2008/De08e. special Issue on Tensor Decompositions and Applications
De Lathauwer, L., De Moor, B., Vandewalle, J.: On the best rank-1 and rank-(R1, R2,., RN) approximation of higher-order tensors. SIAM J. Matrix Anal. Appl. 21, 1324–1342 (2000)
Deng, J., Dong, W., Socher, R., Li, L.J., Li, K., Fei-Fei, L.: ImageNet: a large-scale hierarchical image database. 2009 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 248–255 (2009)
Denil, M., Shakibi, B., Dinh, L., Ranzato, M., de Freitas, N.: Predicting parameters in deep learning. In: Proceedings of the 26th International Conference on Neural Information Processing Systems - Volume 2, NIPS 2013, pp. 2148–2156. Curran Associates Inc. (2013)
Denton, E.L., Zaremba, W., Bruna, J., LeCun, Y., Fergus, R.: Exploiting linear structure within convolutional networks for efficient evaluation. In: Advances in Neural Information Processing Systems, vol. 27, pp. 1269–1277. Curran Associates, Inc. (2014)
Espig, M., Hackbusch, W., Handschuh, S., Schneider, R.: Optimization problems in contracted tensor networks. Comput. Vis. Sci. 14(6), 271–285 (2011)
Figurnov, M., Ibraimova, A., Vetrov, D.P., Kohli, P.: PerforatedCNNs: acceleration through elimination of redundant convolutions. In: Advances in Neural Information Processing Systems, pp. 947–955 (2016)
Gao, X., Zhao, Y., Dudziak, Ł., Mullins, R., Xu, C.Z.: Dynamic channel pruning: feature boosting and suppression. In: International Conference on Learning Representations (2019)
Gusak, J., et al.: Automated multi-stage compression of neural networks. In: 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW), pp. 2501–2508 (2019)
Han, S., Pool, J., Tran, J., Dally, W.: Learning both weights and connections for efficient neural network. In: Cortes, C., Lawrence, N.D., Lee, D.D., Sugiyama, M., Garnett, R. (eds.) Advances in Neural Information Processing Systems, vol. 28, pp. 1135–1143 (2015)
Handschuh, S.: Numerical Methods in Tensor Networks. Ph.D. thesis, Faculty of Mathematics and Informatics, University Leipzig, Germany, Leipzig, Germany (2015)
Harshman, R.A.: Foundations of the PARAFAC procedure: models and conditions for an “explanatory” multimodal factor analysis. In: UCLA Working Papers in Phonetics, vol. 16 pp. 1–84 (1970)
Harshman, R.A.: The problem and nature of degenerate solutions or decompositions of 3-way arrays. In: Tensor Decomposition Workshop, Palo Alto, CA (2004)
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770–778 (2016)
He, Y., Kang, G., Dong, X., Fu, Y., Yang, Y.: Soft filter pruning for accelerating deep convolutional neural networks. In: Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence, IJCAI 2018, pp. 2234–2240 (7 2018)
He, Y., Lin, J., Liu, Z., Wang, H., Li, L.-J., Han, S.: AMC: AutoML for model compression and acceleration on mobile devices. In: Ferrari, V., Hebert, M., Sminchisescu, C., Weiss, Y. (eds.) ECCV 2018. LNCS, vol. 11211, pp. 815–832. Springer, Cham (2018). https://doi.org/10.1007/978-3-030-01234-2_48
Hillar, C.J., Lim, L.H.: Most tensor problems are NP-hard. J. ACM (JACM) 60(6), 45 (2013)
Howard, A., et al.: Searching for MobileNetv3. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 1314–1324 (2019)
Hua, W., Zhou, Y., De Sa, C.M., Zhang, Z., Suh, G.E.: Channel gating neural networks. In: Wallach, H., Larochelle, H., Beygelzimer, A., d’Alché Buc, F., Fox, E., Garnett, R. (eds.) Advances in Neural Information Processing Systems, vol. 32, pp. 1886–1896 (2019)
Khoromskij, B.: \(O(d \log N) \)-quantics approximation of \(N\)-\(d\) tensors in high-dimensional numerical modeling. Constr. Approximation 34(2), 257–280 (2011)
Kim, Y., Park, E., Yoo, S., Choi, T., Yang, L., Shin, D.: Compression of deep convolutional neural networks for fast and low power mobile applications. In: 4th International Conference on Learning Representations, ICLR 2016, San Juan, Puerto Rico, 2–4 May 2016, Conference Track Proceedings (2016). http://arxiv.org/abs/1511.06530
Kossaifi, J., Bulat, A., Tzimiropoulos, G., Pantic, M.: T-net: parametrizing fully convolutional nets with a single high-order tensor. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 7822–7831 (2019)
Kossaifi, J., Toisoul, A., Bulat, A., Panagakis, Y., Hospedales, T.M., Pantic, M.: Factorized higher-order CNNs with an application to spatio-temporal emotion estimation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 6060–6069 (2020)
Krijnen, W., Dijkstra, T., Stegeman, A.: On the non-existence of optimal solutions and the occurrence of “degeneracy” in the CANDECOMP/PARAFAC model. Psychometrika 73, 431–439 (2008)
Krizhevsky, A.: Learning multiple layers of features from tiny images. Technical Report TR-2009, University of Toronto, Toronto (2009)
Landsberg, J.M.: Tensors: Geometry and Applications, vol. 128. American Mathematical Society, Providence (2012)
Lebedev, V.: Algorithms for speeding up convolutional neural networks. Ph.D. thesis, Skoltech, Russia (2018). https://www.skoltech.ru/app/data/uploads/2018/10/Thesis-Final.pdf
Lebedev, V., Ganin, Y., Rakhuba, M., Oseledets, I., Lempitsky, V.: Speeding-up convolutional neural networks using fine-tuned CP-decomposition. In: International Conference on Learning Representations (2015)
Lim, L.H., Comon, P.: Nonnegative approximations of nonnegative tensors. J. Chemom. 23(7–8), 432–441 (2009)
Mitchell, B.C., Burdick, D.S.: Slowly converging PARAFAC sequences: Swamps and two-factor degeneracies. J. Chemom. 8, 155–168 (1994)
Molchanov, D., Ashukha, A., Vetrov, D.: Variational dropout sparsifies deep neural networks. In: Proceedings of the 34th International Conference on Machine Learning - Volume 70, pp. 2498–2507 (2017). JMLR.org
Novikov, A., Podoprikhin, D., Osokin, A., Vetrov, D.: Tensorizing neural networks. In: Proceedings of the 28th International Conference on Neural Information Processing Systems - Volume 1, NIPS 2015, pp. 442–450. MIT Press, Cambridge (2015)
Oseledets, I., Tyrtyshnikov, E.: Breaking the curse of dimensionality, or how to use SVD in many dimensions. SIAM J. Sci. Comput. 31(5), 3744–3759 (2009)
Paatero, P.: Construction and analysis of degenerate PARAFAC models. J. Chemometrics 14(3), 285–299 (2000)
Phan, A.H., Cichocki, A., Uschmajew, A., Tichavský, P., Luta, G., Mandic, D.: Tensor networks for latent variable analysis: novel algorithms for tensor train approximation. IEEE Trans. Neural Network Learn. Syst. (2020). https://doi.org/10.1109/TNNLS.2019.2956926
Phan, A.H., Tichavský, P., Cichocki, A.: Tensor deflation for CANDECOMP/PARAFAC. Part 1: alternating subspace update algorithm. IEEE Trans. Sig. Process. 63(12), 5924–5938 (2015)
Phan, A.H., Tichavský, P., Cichocki, A.: Error preserving correction: a method for CP decomposition at a target error bound. IEEE Trans. Signal Process. 67(5), 1175–1190 (2019)
Phan, A.H., Yamagishi, M., Mandic, D., Cichocki, A.: Quadratic programming over ellipsoids with applications to constrained linear regression and tensor decomposition. Neural Comput. Appl. (2020). https://doi.org/10.1007/s00521-019-04191-z
Rastegari, M., Ordonez, V., Redmon, J., Farhadi, A.: XNOR-Net: ImageNet classification using binary convolutional neural networks. In: Leibe, B., Matas, J., Sebe, N., Welling, M. (eds.) ECCV 2016. LNCS, vol. 9908, pp. 525–542. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-46493-0_32
Rayens, W., Mitchell, B.: Two-factor degeneracies and a stabilization of PARAFAC. Chemometr. Intell. Lab. Syst. 38(2), 173–181 (1997)
Rigamonti, R., Sironi, A., Lepetit, V., Fua, P.: Learning separable filters. In: Proceedings of the 2013 IEEE Conference on Computer Vision and Pattern Recognition. pp. 2754–2761. CVPR ’13, IEEE Computer Society, Washington, DC, USA (2013)
de Silva, V., Lim, L.H.: Tensor rank and the ill-posedness of the best low-rank approximation problem. SIAM J. Matrix Anal. Appl. 30, 1084–1127 (2008)
Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recognition. In: 3rd International Conference on Learning Representations, ICLR (2015)
Stegeman, A., Comon, P.: Subtracting a best rank-1 approximation may increase tensor rank. Linear Algebra Appl. 433(7), 1276–1300 (2010)
Tan, M., Le, Q.V.: EfficientNet: rethinking model scaling for convolutional neural networks. In: ICML (2019)
Tichavský, P., Phan, A.H., Cichocki, A.: Sensitivity in tensor decomposition. IEEE Signal Process. Lett. 26(11), 1653–1657 (2019)
Tucker, L.R.: Implications of factor analysis of three-way matrices for measurement of change. Probl. Measuring Change 15, 122–137 (1963)
Vasilescu, M.A.O., Terzopoulos, D.: Multilinear subspace analysis of image ensembles. In: 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2003), Madison, WI, USA, 16–22 June 2003, pp. 93–99. IEEE Computer Society (2003). https://doi.org/10.1109/CVPR.2003.1211457
Vervliet, N., Debals, O., Sorber, L., Barel, M.V., Lathauwer, L.D.: Tensorlab 3.0, March 2016. http://www.tensorlab.net
Wang, D., Zhao, G., Li, G., Deng, L., Wu, Y.: Lossless compression for 3DCNNs based on tensor train decomposition. CoRR abs/1912.03647 (2019). http://arxiv.org/abs/1912.03647
Zacharov, I., et al.: Zhores – petaflops supercomputer for data-driven modeling, machine learning and artificial intelligence installed in Skolkovo Institute of Science and Technology. Open Eng. 9(1) (2019)
Zhang, T., Golub, G.H.: Rank-one approximation to high order tensors. SIAM J. Matrix Anal. Appl. 23(2), 534–550 (2001). https://doi.org/10.1137/S0895479899352045
Zhang, X., Zou, J., He, K., Sun, J.: Accelerating very deep convolutional networks for classification and detection. IEEE Trans. Pattern Anal. Mach. Intell. 38(10), 1943–1955 (2016)
Zhuang, Z., et al.: Discrimination-aware channel pruning for deep neural networks. In: Advances in Neural Information Processing Systems, pp. 883–894 (2018)
Acknowledgements
The work of A.-H. Phan, A. Cichocki, I. Oseledets, J. Gusak, K. Sobolev, K. Sozykin and D. Ermilov was supported by the Ministry of Education and Science of the Russian Federation under Grant 14.756.31.0001. The results of this work were achieved during the cooperation project with Noah’s Ark Lab, Huawei Technologies. The authors sincerely thank the Referees for very constructive comments which helped to improve the quality and presentation of the paper. The computing for this project was performed on the Zhores CDISE HPC cluster at Skoltech [56].
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this paper
Cite this paper
Phan, AH. et al. (2020). Stable Low-Rank Tensor Decomposition for Compression of Convolutional Neural Network. In: Vedaldi, A., Bischof, H., Brox, T., Frahm, JM. (eds) Computer Vision – ECCV 2020. ECCV 2020. Lecture Notes in Computer Science(), vol 12374. Springer, Cham. https://doi.org/10.1007/978-3-030-58526-6_31
Download citation
DOI: https://doi.org/10.1007/978-3-030-58526-6_31
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-58525-9
Online ISBN: 978-3-030-58526-6
eBook Packages: Computer ScienceComputer Science (R0)