A Study of Aggregation of Long Time-series Input for LSTM Neural Networks
Published in OpenReview / ICLR, 2022
Overview
LSTM networks can struggle to retain information over long input sequences. This paper evaluates an exponential aggregation method where older data is grouped into larger time windows, preserving recent detail while reducing sequence length for the model.
Read on OpenReview Download PDF Google Scholar
Recommended citation: Nitzan Farhi, Yuval Shavitt, "A Study of Aggregation of Long Time-series Input for LSTM Neural Networks." OpenReview, 2022.
Download Paper