Analog circuit integration of backpropagation learning in memristive HTM architecture

Olga Krestinskaya, Alex James

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review


Hierarchical Temporal Memory (HTM) is a biologically plausible model of the neocortex that mimics its structure and functionality. The concepts of HTM and sparse distributed patterns produced by the HTM Spatial Pooler can be useful for various applications. This chapter covers the integration of an analog backpropagation learning circuit into memristive HTM hardware for the Spatial Pooler, which is used for extraction of the meaningful features from the input patterns and conversion of these patterns into a sparse representation followed by the backpropagation learning for the pattern matching based classification. In this chapter, the way to integrate analog learning circuits to the memristive HTM is shown and the advantages of such architecture are discussed. Also, the open research problems in analog memristive HTM systems are summarized.

Original languageEnglish (US)
Title of host publicationMem-elements for Neuromorphic Circuits with Artificial Intelligence Applications
Number of pages12
ISBN (Electronic)9780128211847
StatePublished - Jan 1 2021

Bibliographical note

Publisher Copyright:
© 2021 Elsevier Inc. All rights reserved.


  • Analog circuits
  • Backpropagation learning
  • Hierarchical temporal memory
  • Memristor

ASJC Scopus subject areas

  • General Engineering
  • General Arts and Humanities


Dive into the research topics of 'Analog circuit integration of backpropagation learning in memristive HTM architecture'. Together they form a unique fingerprint.

Cite this