// ⬡ Deep Learning Foundations

Long Short-Term Memory (LSTM)

A gated recurrent network that preserves long-range information.

LSTM 长短期记忆网络 1997 Foundational core

IN PLAIN WORDS

Plain RNN gradients vanish over time. LSTM adds a cell-state conveyor with input/forget/output gates. It ruled sequence modeling before Transformers; xLSTM revisits it.

Used in: Speech recognition, machine translation (2014–2017)

TIMELINE · prerequisites → this → what builds on it

1986 Backpropagati…1997 Long Short-Term M…2006 Connectionist…2014 Sequence-to-S…2014 Gated Recurre…2018 World Models2018 ELMo (Context…

01Original paper

Long Short-Term Memory

Hochreiter et al. · Neural Computation · 1997

DOI ↗ ✓ Verified · Crossref API doi:10.1162/neco.1997.9.8.1735

02Where it sits

Learn the left column first; the right column builds on this one.

Learn first

Backpropagation

Long Short-Term Memory (LSTM)

Then you can learn

Related:Transformer & Self-Attention

Left: prerequisites · right: builds on this · dashed: related. Full map →

03Learning resources

Official docs, explainers, courses and reference code. All links checked.

Explainers

colah：Understanding LSTM Networks ↗EN

Books

Dive into Deep Learning（D2L） ↗EN

04Where the course teaches it

U32 CNNs, RNNs and the Birth of Attention Convolutional networks, LSTMs, Seq2Seq, word vectors and Bahdanau attention. 10 h7 algorithms5 papers

On learning paths:Zero to Deep Learning

For AI / data

Verification: verified, checked 2026-10-21.
JSON: /api/learn/algorithms.json (id: lstm)