paper-with-me

홈 › Papers

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

2021-07-25 · ACL 2021 5 · Zhenhai Zhu, Radu Soricut

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.

📄 PDF Abstract BibTeX arXiv:2107.11906

Code (2)

jinmang2/hierarchical-transformer-1d pytorch
lucidrains/h-transformer-1d pytorch

Tasks

Inductive BiasLanguage Modelling

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Absolute Position Encodings Absolute Position Encodings are a type of position embeddings for [Transformer-based models] where positional encodings are…
Position-Wise Feed-Forward Layer 설명 없음
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
Residual Connection 설명 없음
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…

Similar Papers 제목 키워드 기반

Fast Multipole Attention: A Divide-and-Conquer Attention Mechanism for Long Sequences

2023-10-18 · Yanming Kang, Giang Tran, Hans De Sterck

Transformer-based models have achieved state-of-the-art performance in many areas. However, the quadratic complexity of self-attention with respect to the input length hinders the applicability of Transformer-based model…

Language ModelingLanguage Modelling

Dissociating Decodability and Causal Use in Bracket-Sequence Transformers

2026-04-24 · Aryan Sharma, Cutter Dawes, Shivam Raval arxiv

When trained on tasks requiring an understanding of hierarchical structure, transformers have been found to represent this hierarchy in distinct ways: in the geometry of the residual stream, and in stack-like attention p…

STMT: A Spatial-Temporal Mesh Transformer for MoCap-Based Action Recognition

2023-03-31 · CVPR 2023 1 · Xiaoyu Zhu, Po-Yao Huang, Junwei Liang, Celso M. de Melo 외

We study the problem of human action recognition using motion capture (MoCap) sequences. Unlike existing techniques that take multiple manual steps to derive standardized skeleton representations as model input, we propo…

Action RecognitionTemporal Action Localization

Hierarchical Learning for Generation with Long Source Sequences

2021-04-15 · Tobias Rohde, Xiaoxia Wu, Yinhan Liu

One of the challenges for current sequence to sequence (seq2seq) models is processing long sequences, such as those in summarization and document level machine translation tasks. These tasks require the model to reason a…

DecoderDocument Level Machine TranslationDocument SummarizationDocument Translation+6

Perceiving Longer Sequences With Bi-Directional Cross-Attention Transformers

2024-02-19 · Markus Hiller, Krista A. Ehinger, Tom Drummond

We present a novel bi-directional Transformer architecture (BiXT) which scales linearly with input size in terms of computational cost and memory consumption, but does not suffer the drop in performance or limitation to …

image-classificationImage ClassificationImage SegmentationRetrieval+2