Element-wise Attention Is All You Need
Fuente:
arXiv
Saved in:
| Main Author: | |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910779543912448 |
|---|---|
| author | Feng, Guoxin |
| author_facet | Feng, Guoxin |
| contents | The self-attention (SA) mechanism has demonstrated superior performance across various domains, yet it suffers from substantial complexity during both training and inference. The next-generation architecture, aiming at retaining the competitive performance of SA while achieving low-cost inference and efficient long-sequence training, primarily focuses on three approaches: linear attention, linear RNNs, and state space models. Although these approaches achieve reduced complexity than SA, they all have built-in performance degradation factors, such as diminished “spikiness†and compression of historical information. In contrast to these approaches, we propose a novel element-wise attention mechanism, which uses the element-wise squared Euclidean distance, instead of the dot product operation, to compute similarity and approximates the quadratic complexity term $\exp(q_{ic}k_{jc})$ with a Taylor polynomial. This design achieves remarkable efficiency: during training, the element-wise attention has a complexity of $\mathcal{O}(tLD)$, making long-sequence training both computationally and memory efficient, where $L$ is the sequence length, $D$ is the feature dimension, and $t$ is the highest order of the polynomial; during inference, it can be reformulated as recurrent neural networks, achieving a inference complexity of $\mathcal{O}(tD)$. Furthermore, the element-wise attention circumvents the performance degradation factors present in these approaches and achieves performance comparable to SA in both causal and non-causal forms. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_05730 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Element-wise Attention Is All You Need Feng, Guoxin Machine Learning Artificial Intelligence The self-attention (SA) mechanism has demonstrated superior performance across various domains, yet it suffers from substantial complexity during both training and inference. The next-generation architecture, aiming at retaining the competitive performance of SA while achieving low-cost inference and efficient long-sequence training, primarily focuses on three approaches: linear attention, linear RNNs, and state space models. Although these approaches achieve reduced complexity than SA, they all have built-in performance degradation factors, such as diminished “spikiness†and compression of historical information. In contrast to these approaches, we propose a novel element-wise attention mechanism, which uses the element-wise squared Euclidean distance, instead of the dot product operation, to compute similarity and approximates the quadratic complexity term $\exp(q_{ic}k_{jc})$ with a Taylor polynomial. This design achieves remarkable efficiency: during training, the element-wise attention has a complexity of $\mathcal{O}(tLD)$, making long-sequence training both computationally and memory efficient, where $L$ is the sequence length, $D$ is the feature dimension, and $t$ is the highest order of the polynomial; during inference, it can be reformulated as recurrent neural networks, achieving a inference complexity of $\mathcal{O}(tD)$. Furthermore, the element-wise attention circumvents the performance degradation factors present in these approaches and achieves performance comparable to SA in both causal and non-causal forms. |
| title | Element-wise Attention Is All You Need |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2501.05730 |