We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: H-Transformer-1D: Fast One-Dimensional Hierarchical Attention for Sequences

Abstract: We describe an efficient hierarchical method to compute attention in the Transformer architecture. The proposed attention mechanism exploits a matrix structure similar to the Hierarchical Matrix (H-Matrix) developed by the numerical analysis community, and has linear run time and memory complexity. We perform extensive experiments to show that the inductive bias embodied by our hierarchical attention is effective in capturing the hierarchical structure in the sequences typical for natural language and vision tasks. Our method is superior to alternative sub-quadratic proposals by over +6 points on average on the Long Range Arena benchmark. It also sets a new SOTA test perplexity on One-Billion Word dataset with 5x fewer model parameters than that of the previous-best Transformer-based models.
Comments: ACL2021 long paper oral presentation
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Cite as: arXiv:2107.11906 [cs.LG]
  (or arXiv:2107.11906v1 [cs.LG] for this version)

Submission history

From: Zhenhai Zhu [view email]
[v1] Sun, 25 Jul 2021 23:07:03 GMT (85kb)

Link back to: arXiv, form interface, contact.