Speeding up Transformer Decoding via an Attention Refinement Network (2022.coling-1)
Copied to clipboard
| Challenge: | Extensive experiments on ten WMT machine translation tasks show that the proposed model yields an average of 1.35x faster (with almost no decrease in BLEU) |
| Approach: | They propose a weighted residual network which reconstructs attention by reusing the features across layers. |
| Outcome: | The proposed model is 1.35x faster than the state-of-the-art inference model on translation tasks compared to AAN and SAN models with fewer parameter numbers . |
Similar Papers
Accelerating Neural Transformer via an Average Attention Network (P18-1)
Copied to clipboard
| Challenge: | Using parallelizable attention networks, the neural Transformer is slow to train due to auto-regressive architecture and self-attention in the decoder. |
| Approach: | They propose an average attention network to replace the original self-attention model in the decoder of the neural Transformer. |
| Outcome: | The proposed network can decode sentences over four times faster than the original version with almost no loss in training time and translation performance. |
How Much Attention Do You Need? A Granular Analysis of Neural Machine Translation Architectures (P18-1)
Copied to clipboard
| Challenge: | Neural Machine Translation (NMT) has been replaced by convolutional or self-attentional approaches. |
| Approach: | They propose an architecture definition language that allows for a flexible combination of common building blocks. |
| Outcome: | The proposed architectures can bring recurrent and convolutional models close to the Transformer architecture, but not using self-attention. |
Attention Calibration for Transformer in Neural Machine Translation (2021.acl-long)
Copied to clipboard
| Challenge: | Attention mechanisms have been ubiquitous in neural machine translation (NMT) however, many studies doubt whether highlyattended inputs have a large impact on the model outputs. |
| Approach: | They propose to introduce a mask perturbation model that automatically evaluates each input’s contribution to the model outputs. |
| Outcome: | The proposed model is more uniform at lower layers while more concentrated on the specific inputs at higher layers. |
Mask Attention Networks: Rethinking and Strengthen Transformer (2021.naacl-main)
Copied to clipboard
Zhihao Fan, Yeyun Gong, Dayiheng Liu, Zhongyu Wei, Siyuan Wang, Jian Jiao, Nan Duan, Ruofei Zhang, Xuanjing Huang
| Challenge: | Existing research explores to enhance the two sublayers separately to improve the capability of Transformer for text representation. |
| Approach: | They propose to combine SAN and Feed-Forward Networks to create a dynamic mask attention network with a learnable mask matrix which can model localness adaptively. |
| Outcome: | The proposed model outperforms the original Transformer on translation and text summarization tasks. |
ClusterFormer: Neural Clustering Attention for Efficient and Effective Transformer (2022.acl-long)
Copied to clipboard
| Challenge: | Existing sparse attention methods use fixed patterns to select words without considering similarities between words. |
| Approach: | They propose a neural clustering method which integrates into the Self-Attention Mechanism in Transformer and integrates it into the target task. |
| Outcome: | The proposed method outperforms two typical sparse attention methods on translation, text classification, and text matching tasks while having a comparable or even better time and memory efficiency. |
RealFormer: Transformer Likes Residual Attention (2021.findings-acl)
Copied to clipboard
| Challenge: | Existing techniques to create Residual Attention Layer Transformer networks outperform the canonical Transformer on a wide spectrum of tasks. |
| Approach: | They propose a technique to create Residual Attention Layer Transformer networks that outperform the canonical Transformer on a wide spectrum of tasks. |
| Outcome: | The proposed technique outperforms the canonical Transformer on a wide spectrum of tasks including Masked Language Modeling, GLUE, SQUAD, Neural Machine Translation, WikiHop, HotpotQA, Natural Questions, and OpenKP. |
Recurrent Attention for Neural Machine Translation (2021.emnlp-main)
Copied to clipboard
| Challenge: | Recent research questions the importance of dot-product self-attention in Transformer models and shows that most attention heads learn simple positional patterns. |
| Approach: | They propose a novel mechanism to replace dot-product self-attention with a recurrent atteNtion mechanism that directly learns attention weights without token-to-token interaction. |
| Outcome: | The proposed model outperforms the Transformer model on translation tasks with fewer parameters and inference time. |
Synchronous Syntactic Attention for Transformer Neural Machine Translation (2021.acl-srw)
Copied to clipboard
| Challenge: | Existing syntaxbased NMT models use monolingual syntactic information on either side or both. |
| Approach: | They propose a mechanism that synchronizes source-side and target-side syntactic self-attentions by minimizing the difference between target- and target side self- attentions mapped by the encoder-decoder attention matrix. |
| Outcome: | The proposed method improves translation performance on WMT14 En-De, WMT16 En-Ro, and ASPEC Ja-En (up to +0.38 points in BLEU). |
Learning Hard Retrieval Decoder Attention for Transformers (2021.findings-emnlp)
Copied to clipboard
| Challenge: | In this paper, we show that learning a hard retrieval attention that attends to a single token in a sentence is 1.43 times faster than the standard scaled dot-product attention. |
| Approach: | They propose a method to learn hard retrieval attention where an attention head attends to a single token in a sentence rather than all tokens. |
| Outcome: | The proposed method is 1.43 times faster in decoding while preserving translation quality on a wide range of MT tasks. |
Improving Deep Transformer with Depth-Scaled Initialization and Merged Attention (D19-1)
Copied to clipboard
| Challenge: | Existing methods to improve NLP convergence and computational overhead are limited by stacking more layers. |
| Approach: | They propose a depth-scaled initialization method which reduces parameter variance at initialization and reduces output variance of residual connections to ease gradient back-propagation. |
| Outcome: | The proposed method outperforms the base model on translation tasks with five translation directions while matching the decoding speed of the baseline model. |