Advertisement

Scalable Prototype Learning Using GPUs

  • Tonghua SuEmail author
  • Songze Li
  • Peijun Ma
  • Shengchun Deng
  • Guangsheng Liang
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8814)

Abstract

Prototype learning is widely used in character recognition field. Unfortunately, current learning algorithms require intensive computation burden for large category applications, such as Japanese/Chinese character recognition. To resolve this challenge, a principled parallel method is proposed on GPUs instead of CPUs. We have implemented the method in mini-batch manner as well as stochastic gradient descent (SGD) manner. Our evaluations on a Chinese character database show that our method posses a high scalability while preserving its performance precision. Up to 194X speedup can be achieved in the case of mini-batch. Even to the more difficult SGD occasion, a more than 30-fold speedup is observed.

Keywords

Prototype learning Learning vector quantization Chinese character recognition Parallel reduction GPU computing CUDA 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Liu, C., Nakagawa, M.: Evaluation of prototype learning algorithms for nearest-neighbor classifier in application to handwritten character recognition. Pattern Recognition 34, 601–615 (2001)CrossRefzbMATHGoogle Scholar
  2. Su, T., Liu, C., Zhang, X.: Perceptron learning of modified quadratic discriminant function. In: ICDAR (2011)Google Scholar
  3. Lv, Y., Huang, L., et al.: Learning-Based Candidate Segmentation Scoring for Real-Time Recognition of Online Overlaid Chinese Handwriting. In: ICDAR (2013)Google Scholar
  4. Su, T., Ma, P., et al.: Exploring MPE/MWE Training for Chinese Handwriting Recognition. In: ICDAR (2013)Google Scholar
  5. Zhou, M., Yin, F., Liu, C.: GPU-Based Fast Training of Discriminative Learning Quadratic Discriminant Function for Handwritten Chinese Character Recognition. In: ICDAR (2013)Google Scholar
  6. Raina, R., Madhavan, A., Ng, A.Y.: Large-scale deep unsupervised learning using graphics processors. In: ICML (2009)Google Scholar
  7. Ciresan, D.C., Meier, U., et al.: Deep, big, simple neural nets for handwritten digit recognition. Neural Computation 22, 3207–3220 (2010)CrossRefGoogle Scholar
  8. Sato, A., Yamada, K.: Generalized learning vector quantization. In: NIPS (1996)Google Scholar
  9. Wilt, N.: The CUDA Handbook: A Comprehensive Guide to GPU Programming. Addison-Wesley (2013)Google Scholar
  10. Kirk, D.B., Wen-mei, W.H.: Programming massively parallel processors: a hands-on approach. Morgan Kaufmann (2010)Google Scholar
  11. Liu, C., Yin, F., et al.: CASIA Online and Offline Chinese Handwriting Databases. In: ICDAR (2011)Google Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  • Tonghua Su
    • 1
    Email author
  • Songze Li
    • 1
  • Peijun Ma
    • 1
  • Shengchun Deng
    • 1
  • Guangsheng Liang
    • 2
  1. 1.School of SoftwareHarbin Institute of TechnologyHarbinPeople’s Republic of China
  2. 2.The Second Artillery Force of the PLABeijingPeople’s Republic of China

Personalised recommendations