Training recurrent memory in real time

Real-time recurrent learning was applied to recognizing sequential structure in finite-state grammars.

Anthony W. Smith and David Zipser

AI topics

Explore related entries. Larger tags appear on more entries.

The contribution

Smith and Zipser examined learning in networks with recurrent connections, where retained activity lets earlier events affect current computations. Their study applied the real-time recurrent learning algorithm to sequential structure, contributing to the line of research on how a network can learn to use its own evolving state as memory.

What this does not establish

Real-time refers to the way learning updates are computed, not a guarantee of cheap computation or perfect memory. Grammar-learning experiments do not establish general reasoning ability.

Why this date?

1989 dates the journal article. The original title uses the singular Structure; LSTM's bibliography uses the plural Structures for the same authors, venue, and page range.

This entry follows the linked publication. Read the source and date conventions.

The original work

Learning Sequential Structure with the Real-Time Recurrent Learning Algorithm

Authors: Anthony W. Smith and David Zipser
Publication: International Journal of Neural Systems, 1 (2), 125-131
Source type: Journal article

Original journal record

The links above support the description and dating of this entry. Full text may be open or publisher-restricted.

Cited By

Atlas entries whose reference lists cite this work or the publication version noted below.

Long Short-Term Memory (1997)
Cites Smith and Zipser (1989), International Journal of Neural Systems 1 (2), 125-131. Check the reference list.

Citations

Works in the Atlas cited by this entry.

No references to other Atlas entries have been recorded for this work yet.

Bibliography coverage: Bibliography not yet available. Open the source.

The publisher record supplies no reference list through Crossref. After human verification, the publisher page was inspected: its References tab says None and the article is marked No Access. The abstract contains numbered citations, so the empty HTML reference tab does not mean no bibliography. Alternate repository and author-copy searches have not recovered the seven-page paper. The original reference page is still required.

Download the bibliography index · Download Atlas citation links

These links cover verified matches within the Atlas. They are not total scholarly citation counts. A citation alone does not establish that the later system implements the same method.

AI topics

The contribution

What this does not establish

Why this date?

Comments