Abstract
Stochastic Gradient Descent (SGD) is widely used to train a machine learning model over large datasets, yet its slow convergence rate can be a bottleneck. As a remarkable family of variance reduction techniques, memory algorithms such as SAG and SAGA have been proposed to accelerate the convergence rate of SGD. However, these algorithms need to store per training data point corrections in memory. The unlimited space usage feature is impractical for modern large-scale applications, especially over data points that arrive over time (referred to as streaming data in this paper). To overcome this weakness, this paper investigates the methods that bound the space usage in the state-of-the-art family of variance-reduced stochastic gradient descent over streaming data, and presents CHEAPS2AGA. At each step of updating the model, the key idea of CHEAPS2AGA is always reserving N random data points as samples, while re-using information about past stochastic gradients across all the observed data points with limited space usage. In addition, training an accurate model over streaming data requires the algorithm to be time-efficient. To accelerate the model training phase, CHEAPS2AGA embraces a lock-free data structure to insert new data points and remove unused data points in parallel, and updates the model parameters without using any locking. We conduct comprehensive experiments to compare CHEAPS2AGA to prior related algorithms suited for streaming data. The experimental results demonstrate the practical competitiveness of CHEAPS2AGA in terms of scalability and accuracy.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
FAA(addr, v): assuming the content stored in addr are a, this operation returns a and stores \(a+v\) at addr.
- 2.
SWAP(addr, v): assuming the content stored in addr are a, this operation returns a and stores v at addr.
References
Bertsekas, D.P.: Nonlinear Programming, 3rd edn. Athena Scientific, Nashua (2016)
Bottou, L., Cun, Y.L.: Large scale online learning. In: Advances in Neural Information Processing Systems, vol. 16, pp. 217–224 (2004)
Daneshmand, H., Lucchi, A., Hofmann, T.: Starting small: learning with adaptive sample sizes. In: Proceedings of the 33rd International Conference on International Conference on Machine Learning, ICML 2016, pp. 1463–1471 (2016)
Defazio, A., Bach, F., Lacoste-Julien, S.: SAGA: a fast incremental gradient method with support for non-strongly convex composite objectives. In: Proceedings of the 27th International Conference on Neural Information Processing Systems, NIPS 2014, pp. 1646–1654 (2014)
Frostig, R., Rong, G., Kakade, S.M., Sidford, A.: Competing with the empirical risk minimizer in a single pass. In: Proceedings of the 28th Annual Conference on Learning Theory, COLT 2015, pp. 728–763 (2015)
Hofmann, T., Lucchi, A., Lacoste-Julien, S., McWilliams, B.: Variance reduced stochastic gradient descent with neighbors. In: Proceedings of the 28th International Conference on Neural Information Processing Systems, NIPS 2015, vol. 2, pp. 2305–2313. MIT Press, Cambridge (2015)
Reddi, S.J., Hefny, A., Sra, S., Poczos, B., Smola, A.J.: On variance reduction in stochastic gradient descent and its asynchronous variants. In: Advances in Neural Information Processing Systems, vol. 28, pp. 2647–2655 (2015)
Johnson, R., Zhang, T.: Accelerating stochastic gradient descent using predictive variance reduction. In: Proceedings of the 26th International Conference on Neural Information Processing Systems, NIPS 2013, pp. 315–323 (2013)
Jothimurugesan, E., Tahmasbi, A., Gibbons, P., Tirthapura, S.: Variance-reduced stochastic gradient descent on streaming data. In: Advances in Neural Information Processing Systems, vol. 31, pp. 9906–9915 (2018)
Konećný, J., Richtárik, P.: Semi-stochastic gradient descent methods. arXiv:1312.1666 (2013)
Leblond, R., Pedregosa, F., Lacoste-Julien, S.: ASAGA: asynchronous parallel SAGA. In: Proceedings of the 20th International Conference on Artificial Intelligence and Statistics, AISTATS 2017, pp. 46–54 (2017)
Lee, J., et al.: Wide neural networks of any depth evolve as linear models under gradient descent. In: Advances in Neural Information Processing Systems, vol. 32, pp. 8572–8583. Curran Associates, Inc. (2019)
Lewis, D.D., Yang, Y., Rose, T.G., Li, F.: RCV1: a new benchmark collection for text categorization research. J. Mach. Learn. Res. 5, 361–397 (2004)
Ma, J., Saul, L.K., Savage, S., Voelker, G.M.: Identifying suspicious URLs: an application of large-scale online learning. In: Proceedings of the 26th Annual International Conference on Machine Learning, ICML 2009, pp. 681–688 (2009)
Peng, Y., Hao, Z., Yun, X.: Lock-free parallelization for variance-reduced stochastic gradient descent on streaming data. IEEE Trans. Parallel Distrib. Syst. 31(9), 2220–2231 (2020)
Recht, B., Re, C., Wright, S., Niu, F.: HOGWILD!: a lock-free approach to parallelizing stochastic gradient descent. In: Advances in Neural Information Processing Systems, vol. 24, pp. 693–701 (2011)
Robbins, H., Monro, S.: A stochastic approximation method. Ann. Math. Stat. 22(3), 400–407 (1951)
Roux, N.L., Schmidt, M., Bach, F.: A stochastic gradient method with an exponential convergence rate for finite training sets. In: Proceedings of the 25th International Conference on Neural Information Processing Systems, NIPS 2012, vol. 2, pp. 2663–2671 (2012)
Shah, V., Asteris, M., Kyrillidis, A., Sanghavi, S.: Trading-off variance and complexity in stochastic gradient descent. arXiv:1603.06861 (2016)
Zhang, H., Hsieh, C., Akella, V.: HOGWILD++: a new mechanism for decentralized asynchronous stochastic gradient descent. In: Proceedings of the 16th International Conference on Data Mining, ICDM 2016, pp. 629–638 (2016)
Zhao, P., Chen, P., Wang, S., Lin, X.: Towards query-efficient black-box adversary with zeroth-order natural gradient descent. In: Proceedings of the 34th AAAI Conference on Artificial Intelligence, AAAI 2020, pp. 6909–6916 (2020)
Acknowledgements
This work was supported by National Natural Science Foundation of China under grant No.61702499. The authors would like to thank the anonymous reviewers for their useful suggestions which helped to improve the quality of the paper.
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
Peng, Y., Fei, H., Li, L., Ding, Z., Hao, Z. (2020). CHEAPS2AGA: Bounding Space Usage in Variance-Reduced Stochastic Gradient Descent over Streaming Data and Its Asynchronous Parallel Variants. In: Qiu, M. (eds) Algorithms and Architectures for Parallel Processing. ICA3PP 2020. Lecture Notes in Computer Science(), vol 12453. Springer, Cham. https://doi.org/10.1007/978-3-030-60239-0_27
Download citation
DOI: https://doi.org/10.1007/978-3-030-60239-0_27
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-60238-3
Online ISBN: 978-3-030-60239-0
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)