Long Context & Efficient Sequences · 2020

Linearized Attention (Reformer/Performer/Linear Transformers)

Nikita Kitaev, Łukasz Kaiser, Krzysztof Choromanski, Adrian Weller, Angelos Katharopoulos, François Fleuret

Linearized-attention models replaced the softmax attention matrix with kernel feature maps or hashing so attention could be computed in linear time and memory, with Performer using random features (FAVOR+) to approximate softmax attention without ever forming the full N-by-N matrix.

Editorial record

Plain-language summary

Softmax attention requires materializing an N-by-N similarity matrix; these methods avoid it. Performer's FAVOR+ maps queries and keys through random feature functions whose dot products approximate the softmax kernel unbiasedly, letting attention be reassociated into linear-time operations. Reformer used locality-sensitive hashing and reversible layers to similar ends, and Linear Transformers recast attention as a kernelized RNN. This traded exact attention for near-linear scaling, enabling much longer sequences on fixed memory, though with some approximation cost that limited adoption at frontier scale.

Knowledge graph

Relationships

Antecedents

Descendants

Source record

Provenance

Record ID
P-421
Record created
2026-07-13
Last reviewed
2026-07-14
Record version
2

Citation caveat: Citation metadata is approximate and marked unverified in the source dataset.