Findings of the Association for Computational Linguistics: EMNLP 2021 2021
DOI: 10.18653/v1/2021.findings-emnlp.39
|View full text |Cite
|
Sign up to set email alerts
|

Attention Weights in Transformer NMT Fail Aligning Words Between Sequences but Largely Explain Model Predictions

Abstract: This work proposes an extensive analysis of the Transformer architecture in the Neural Machine Translation (NMT) setting. Focusing on the encoder-decoder attention mechanism, we prove that attention weights systematically make alignment errors by relying mainly on uninformative tokens from the source sequence. However, we observe that NMT models assign attention to these tokens to regulate the contribution in the prediction of the two contexts, the source and the prefix of the target sequence. We provide evide… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2

Citation Types

0
12
0

Year Published

2022
2022
2024
2024

Publication Types

Select...
4
2

Relationship

1
5

Authors

Journals

citations
Cited by 9 publications
(12 citation statements)
references
References 17 publications
0
12
0
Order By: Relevance
“…To compute the relative contribution of source and target tokens to each prediction made by the system we used an embedding perturbation method [41]. Given a source sentence x and its translation y, the absolute source contribution C S (y j ) when producing the probability of the j-th token y j is defined as the variance of y j 's output probability across N random perturbations of the word embeddings of x.…”
Section: Relative Source and Target Contributionsmentioning
confidence: 99%
See 1 more Smart Citation
“…To compute the relative contribution of source and target tokens to each prediction made by the system we used an embedding perturbation method [41]. Given a source sentence x and its translation y, the absolute source contribution C S (y j ) when producing the probability of the j-th token y j is defined as the variance of y j 's output probability across N random perturbations of the word embeddings of x.…”
Section: Relative Source and Target Contributionsmentioning
confidence: 99%
“…Note that this approach involves replacing some words in targetlanguage sentences with other words randomly chosen from 19. We set N = 50 and λ = 0.01 [41]. For high-resource conditions, mean and standard deviation of the source influence obtained when translating in-domain test sets with the baseline system, four other DA reference systems, and MaTiLDA using different transformations and combinations of them.…”
Section: Relative Source and Target Contributionsmentioning
confidence: 99%
“…former computes gradients with respect to the input token embeddings to measure how much a change in the input changes the output. However, there is a tension between finding a faithful explanation and observing human-like alignments, since one does not imply the other (Ferrando and Costa-jussà, 2021).…”
Section: Introductionmentioning
confidence: 99%
“…Nonetheless, they apply their method on average over a dataset, not for getting input attributions of a single prediction. Gradientbased methods have also been extended to the target prefix (Ferrando and Costa-jussà, 2021), although they do not quantify the relative contribution of source and target inputs.…”
Section: Introductionmentioning
confidence: 99%
“…In ARE-based methods, which rely on an underlying classifier to predict if a post is toxic or not, the classifier is trained on the training part of the fold (which contains only toxic posts, ignoring the toxic span annotations) and a randomly selected but not in toxicity detection. See alsoWiegreffe and Pinter (2019),Kobayashi et al (2020), Ferrando andCosta-jussà (2021) for a broader discussion of attention as an explainability mechanism.…”
mentioning
confidence: 99%