Skip to main content

Long Distance Relationships Without Time Travel: Boosting the Performance of a Sparse Predictive Autoencoder in Sequence Modeling

  • Conference paper
  • First Online:
Artificial Neural Networks in Pattern Recognition (ANNPR 2020)


In sequence learning tasks such as language modelling, Recurrent Neural Networks must learn relationships between input features separated by time. State of the art models such as LSTM and Transformer are trained by backpropagation of losses into prior hidden states and inputs held in memory. This allows gradients to travel through time from present to past and effectively learn with perfect hindsight, but at a significant memory cost. In this paper we show that it is possible to train high performance recurrent networks using information that is local in time, and thereby achieve a significantly reduced memory footprint. We describe a predictive autoencoder called bRSM featuring recurrent connections, sparse activations, and a boosting rule for improved cell utilization. The architecture demonstrates near optimal performance on a non-deterministic (stochastic) partially-observable sequence learning task consisting of high-Markov-order sequences of MNIST digits. We find that this model learns these sequences faster and more completely than an LSTM, and offer several possible explanations why the LSTM architecture might struggle with the partially observable sequence structure in this task. We also apply our model to a next word prediction task on the Penn Treebank (PTB) dataset. We show that a ‘flattened’ RSM network, when paired with a modern semantic word embedding and the addition of boosting, achieves 103.5 PPL (a 20-point improvement over the best N-gram models), beating ordinary RNNs trained with BPTT and approaching the scores of early LSTM implementations. This work provides encouraging evidence that strong results on challenging tasks such as language modeling may be possible using less memory intensive and more biologically-plausible training regimes.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Subscribe and save

Springer+ Basic
EUR 32.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
USD 49.99
Price excludes VAT (USA)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Softcover Book
USD 64.99
Price excludes VAT (USA)
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Similar content being viewed by others


  1. 1.

    Code for the bRSM model and all experiments is available at


  1. Ahmad, S., Scheinkman, L.: How can we be so dense? The benefits of using highly sparse representations. arXiv preprint arXiv:1903.11257 (2019)

  2. Bai, S., Kolter, J.Z., Koltun, V.: An empirical evaluation of generic convolutional and recurrent networks for sequence modeling. arXiv preprint arXiv:1803.01271 (2018)

  3. Belkin, M., Hsu, D., Ma, S., Mandal, S.: Reconciling modern machine learning and the bias-variance trade-off. arXiv preprint arXiv:1812.11118 (2018)

  4. Bojanowski, P., Grave, E., Joulin, A., Mikolov, T.: Enriching word vectors with subword information. Trans. Assoc. Comput. Linguist. 5, 135–146 (2017)

    Article  Google Scholar 

  5. Gong, C., He, D., Tan, X., Qin, T., Wang, L., Liu, T.Y.: FRAGE: frequency-agnostic word representation. In: Advances in Neural Information Processing Systems, pp. 1334–1345 (2018)

    Google Scholar 

  6. Hao, K.: Training a single AI model can emit as much carbon as five cars in their lifetimes (2019).

  7. Hawkins, J., Ahmad, S., Purdy, S., Lavin, A.: Biological and machine intelligence (2016). Initial online release 0.4

  8. Hawkins, J., Ahmad, S.: Why neurons have thousands of synapses, a theory of sequence memory in neocortex. Front. Neural Circ. 10 (2016).

  9. Hochreiter, S., Schmidhuber, J.: Long short-term memory. Neural Comput. 9(8), 1735–1780 (1997).

    Article  Google Scholar 

  10. Makhzani, A., Frey, B.: K-Sparse Autoencoders. arXiv:1312.5663 [cs], December 2013

  11. Marcus, M., et al.: The Penn Treebank: annotating predicate argument structure. In: Proceedings of the workshop on Human Language Technology, pp. 114–119. Association for Computational Linguistics (1994)

    Google Scholar 

  12. Melis, G., Kočiský, T., Blunsom, P.: Mogrifier LSTM. arXiv:1909.01792 (2019)

  13. Mikolov, T., Kombrink, S., Burget, L., Černocký, J., Khudanpur, S.: Extensions of recurrent neural network language model. In: 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 5528–5531, May 2011.

  14. Mikolov, T., Deoras, A., Kombrink, S., Burget, L., Černockỳ, J.: Empirical evaluation and combination of advanced language modeling techniques. In: Twelfth Annual Conference of the International Speech Communication Association (2011)

    Google Scholar 

  15. Mikolov, T., Karafiát, M., Burget, L., Černockỳ, J., Khudanpur, S.: Recurrent neural network based language model. In: Eleventh Annual Conference of the International Speech Communication Association (2010)

    Google Scholar 

  16. van den Oord, A., et al.: WaveNet: a generative model for raw audio. arXiv:1609.03499 [cs], September 2016

  17. Rawlinson, D., Ahmed, A., Kowadlo, G.: Learning distant cause and effect using only local and immediate credit assignment. arXiv:1905.11589 [cs, stat], May 2019

  18. Vaswani, A., et al.: Attention is all you need. In: Advances in Neural Information Processing Systems, pp. 5998–6008 (2017)

    Google Scholar 

  19. Williams, R.J., Peng, J.: An efficient gradient-based algorithm for on-line training of recurrent network trajectories. Neural Comput. 2(4), 490–501 (1990)

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations


Corresponding author

Correspondence to Jeremy Gordon .

Editor information

Editors and Affiliations

Rights and permissions

Reprints and permissions

Copyright information

© 2020 Springer Nature Switzerland AG

About this paper

Check for updates. Verify currency and authenticity via CrossMark

Cite this paper

Gordon, J., Rawlinson, D., Ahmad, S. (2020). Long Distance Relationships Without Time Travel: Boosting the Performance of a Sparse Predictive Autoencoder in Sequence Modeling. In: Schilling, FP., Stadelmann, T. (eds) Artificial Neural Networks in Pattern Recognition. ANNPR 2020. Lecture Notes in Computer Science(), vol 12294. Springer, Cham.

Download citation

  • DOI:

  • Published:

  • Publisher Name: Springer, Cham

  • Print ISBN: 978-3-030-58308-8

  • Online ISBN: 978-3-030-58309-5

  • eBook Packages: Computer ScienceComputer Science (R0)

Publish with us

Policies and ethics