CoChat

Is Attention Interpretable?

Open access
Serrano, Sofia, Smith, Noah A. · 2019 · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics

The DOI matches its Crossref record; 43 of the 43 references it cites were found in a registry or an index.

Abstract

Attention mechanisms have recently boosted performance on a range of NLP tasks.Because attention layers explicitly weight input components' representations, it is also often assumed that attention can be used to identify information that models found important (e.g., specific contextualized word tokens).We test whether that assumption holds by manipulating attention weights in already-trained text classification models and analyzing the resulting differences in their predictions.While we observe some ways in which higher attention weights correlate with greater impact on model predictions, we also find many ways in which this does not hold, i.e., where gradient-based rankings of attention weights better predict their effects than their magnitudes.We conclude that while attention noisily predicts input components' overall importance to a model, it is by no means a fail-safe indicator.1

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.18653/v1/p19-1282

Cites

43 of 43 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Our method for calculating the importance of representations corresponding to zeroed-out attention weights, in a hypothetical setting with four output classes .
Figure 1 Our method for calculating the importance of representations corresponding to zeroed-out attention weights, in a hypothetical setting with four output classes .CC-BY · From the PDF
Image 2
CC-BY · From the PDF
Image 3
CC-BY · From the PDF
Figure 4 These are the counts of test instances for the HANrnn models for which i ∗ 's JS divergence was smaller, binned by ∆ α . These counts comprise a small fraction of the test set sizes listed in Table 1.
Figure 4 These are the counts of test instances for the HANrnn models for which i ∗ 's JS divergence was smaller, binned by ∆ α . These counts comprise a small fraction of the test set sizes listed in Table 1.CC-BY · From the PDF
Table 2: Percent of test instances in each decision-flip indicator variable category for each HANrnn.
Table 2: Percent of test instances in each decision-flip indicator variable category for each HANrnn.CC-BY · From the PDF
Image 6
CC-BY · From the PDF
Image 7
CC-BY · From the PDF
Image 8
CC-BY · From the PDF
Image 9
CC-BY · From the PDF
Image 10
CC-BY · From the PDF
Image 11
CC-BY · From the PDF
Image 12
CC-BY · From the PDF
Difference in JS Divergences from Original Output Distribution
Difference in JS Divergences from Original Output DistributionCC-BY · From the PDF
Image 14
CC-BY · From the PDF
Image 15
CC-BY · From the PDF
Figure 9 Using the definition of i ∗ given in section 4 (the highest-attention-weight attended item) and comparing to a different randomly selected attended item, these were the percentages of test instances that fell into each decision-flip indicator variable category for each of the four test sets on all models. Since we require our random item not to be i ∗ , we exclude any instances with a final sequence length of 1 (one sentence for the HANs, one word for the FLANs) from analysis.
Figure 9 Using the definition of i ∗ given in section 4 (the highest-attention-weight attended item) and comparing to a different randomly selected attended item, these were the percentages of test instances that fell into each decision-flip indicator variable category for each of the four test sets on all models. Since we require our random item not to be i ∗ , we exclude any instances with a final sequence length of 1 (one sentence for the HANs, one word for the FLANs) from analysis.CC-BY · From the PDF
Figure 10 Distribution of fraction of attention weights that had to be removed by different ranking schemes to change each model architecture's decisions for each of the four datasets. The different rankings (aside from 'Attention', which corresponds to the attention weight magnitudes in descending order) are described in section 5.2.
Figure 10 Distribution of fraction of attention weights that had to be removed by different ranking schemes to change each model architecture's decisions for each of the four datasets. The different rankings (aside from 'Attention', which corresponds to the attention weight magnitudes in descending order) are described in section 5.2.CC-BY · From the PDF
Figure 11 Distribution of probability masses that had to be removed by different ranking schemes to change each model architecture's decisions for each of the four datasets. While we do not discuss these in the paper due to space constraints, we notice that in most cases, a high fraction of the original attention distribution's probability mass must be zeroed before the (renormalized) modified attended representation results in a changed decision using the Attention ranking.
Figure 11 Distribution of probability masses that had to be removed by different ranking schemes to change each model architecture's decisions for each of the four datasets. While we do not discuss these in the paper due to space constraints, we notice that in most cases, a high fraction of the original attention distribution's probability mass must be zeroed before the (renormalized) modified attended representation results in a changed decision using the Attention ranking.CC-BY · From the PDF
Figure 12 Let i ∗ g be the highest-ranked attended item using our purely gradient-based ranking of importance described in section 5.2. We rerun our single-weight decision flip tests using this new i ∗ g , comparing to a different randomly selected attended item. These were the percentages of test instances that fell into each decision-flip indicator variable category for each of the four test sets on all models. Since we require our random item not to be i ∗ g , we exclude any instances with a final sequence length of 1 (one sentence for the HANs, one word for the FLANs) from analysis.
Figure 12 Let i ∗ g be the highest-ranked attended item using our purely gradient-based ranking of importance described in section 5.2. We rerun our single-weight decision flip tests using this new i ∗ g , comparing to a different randomly selected attended item. These were the percentages of test instances that fell into each decision-flip indicator variable category for each of the four test sets on all models. Since we require our random item not to be i ∗ g , we exclude any instances with a final sequence length of 1 (one sentence for the HANs, one word for the FLANs) from analysis.CC-BY · From the PDF
Figure 13 Let i ∗ p be the highest-ranked attended item using our attention-gradient product ranking of importance described in section 5.2. Once again, we rerun our single-weight decision flip tests using this new i ∗ p , comparing to a different randomly selected attended item. These were the percentages of test instances that fell into each decision-flip indicator variable category for each of the four test sets on all models. Since we require our random item not to be i ∗ p , we exclude any instances with a final sequence length of 1 (one sentence for the HANs, one word for the FLANs) from analysis.
Figure 13 Let i ∗ p be the highest-ranked attended item using our attention-gradient product ranking of importance described in section 5.2. Once again, we rerun our single-weight decision flip tests using this new i ∗ p , comparing to a different randomly selected attended item. These were the percentages of test instances that fell into each decision-flip indicator variable category for each of the four test sets on all models. Since we require our random item not to be i ∗ p , we exclude any instances with a final sequence length of 1 (one sentence for the HANs, one word for the FLANs) from analysis.CC-BY · From the PDF

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

  • Attention Is All You Need

    Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404

  • Adam: A Method for Stochastic Optimization

    Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · cited 1×

    “Unless otherwise noted, to train each model, we used Adam (Kingma and Ba, 2014) with gradient clipping of 10.0 and a patience value of 5, so we would stop training a model if five epochs elapsed without any improvement in validation set accuracy.”
    MethodInfluential
  • "Why Should I Trust You?": Explaining the Predictions of Any Classifier

    Ribeiro, M. T., Singh, S., & Guestrin, C. (2016) · Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/2939672.2939778 · Cited by 17,555 · cited 4×

    “Note, however, that we still do not take either as a gold-standard indicator of importance to the model, as with the gradient in Ross et al. (2017) and Melis and Jaakkola (2018), but merely as an alternative ordering method.”
    BackgroundMethodInfluential
  • Neural Machine Translation by Jointly Learning to Align and Translate

    Bahdanau, D., Cho, K., & Bengio, Y. (2014) · arXiv · doi:10.48550/arxiv.1409.0473 · Cited by 16,213 · cited 2×

    “Specifically, we use the additive formulation originally defined in Bahdanau et al. (2015).”
    BackgroundMethod
  • Convolutional Neural Networks for Sentence Classification

    Kim, Y. (2014) · Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/v1/d14-1181 · Cited by 14,831 · cited 2×

    “Aside from biRNNs, we also experiment with models that instead contextualize word vectors by convolutions on only a token’s close neighbors, inspired by Kim (2014). See Figure 2 for a diagram of the FLAN architecture using a convolutional encoder.”
    Method
  • Effective Approaches to Attention-based Neural Machine Translation

    Luong, T., Pham, H., & Manning, C. D. (2015) · Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d15-1166 · Cited by 8,382 · cited 1×

    “For many different problems — to name a few, machine translation (Luong et al., 2015), syntactic parsing (Vinyals et al., 2015), reading comprehension (Hermann et al., 2015), and language modeling (Liu and Lapata, 2018) — incorporating attention mechanisms into models has proven beneficial for…”
    Background
  • “Lipton (2016) describes a model as “transparent” if “a person can contemplate the entire model at once.””
    Background
  • Hierarchical Attention Networks for Document Classification

    Yang, Z., Yang, D., Dyer, C., He, X., Smola, A., & Hovy, E. (2016) · Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies · doi:10.18653/v1/n16-1174 · Cited by 4,712 · cited 6×

    “Our model architectures are inspired by the hierarchical attention network (HAN; Yang et al., 2016), a text classification model with two layers of attention, first to the word tokens in each sentence and then to the resulting sentence representations.”
    BackgroundMethodResultInfluential
  • Character-level Convolutional Networks for Text Classification

    Zhang, X., Zhao, J., & LeCun, Y. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1509.01626 · Cited by 3,205 · cited 3×

    “We investigate four model architectures on a topic classification dataset (Yahoo Answers; Zhang et al., 2015) and on three review ratings datasets: IMDB (Diao et al.”
    Method
  • Teaching Machines to Read and Comprehend

    Hermann, K. M., Kočiský, T., Grefenstette, E., Espeholt, L., Kay, W., Suleyman, M., & Blunsom, P. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1506.03340 · Cited by 2,521 · cited 1×

    “For many different problems—to name a few, machine translation (Luong et al., 2015), syntactic parsing (Vinyals et al., 2015), reading comprehension (Hermann et al., 2015), and language modeling (Liu and Lapata, 2018)—incorporating attention mechanisms into models has proven benefi-cial for…”
    Background
  • Attention-based LSTM for Aspect-level Sentiment Classification

    Wang, Y., Huang, M., zhu, X., & Zhao, L. (2016) · Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d16-1058 · Cited by 2,450 · cited 1×

    “Since attention calculates a distribution over inputs, prior work has used attention as a tool for interpretation of model decisions (Wang et al., 2016; Lee et al., 2017; Lin et al., 2017; Ghaeini et al., 2018).”
    Method
  • Attention is not Explanation

    Jain, S., & Wallace, B. (2019) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.18653/v1/n19-1357 · Cited by 1,987 · cited 3×

    “So far, other work looking at attention has examined whether attention suffices as a holistic explanation for a model’s decision (Jain and Wallace, 2019), which is a higher bar.”
    BackgroundResultInfluential
  • A Structured Self-attentive Sentence Embedding

    Lin, Z., Feng, M., Santos, C. N. D., Yu, M., Xiang, B., Zhou, B., & Bengio, Y. (2017) · arXiv · doi:10.48550/arxiv.1703.03130 · Cited by 1,743 · cited 1×

    “Since attention calculates a distribution over inputs, prior work has used attention as a tool for interpretation of model decisions (Wang et al., 2016; Lee et al., 2017; Lin et al., 2017; Ghaeini et al., 2018).”
    Method
  • Understanding Black-box Predictions via Influence Functions

    Koh, P. W., & Liang, P. (2017) · arXiv · doi:10.48550/arxiv.1703.04730 · Cited by 1,514 · cited 2×

    “For example, recent work has focused on which training instances (Koh and Liang, 2017) or which human-interpretable features were most relevant for a particular decision (Ribeiro et al., 2016; Arras et al., 2016).”
    Background
  • Sanity Checks for Saliency Maps

    Adebayo, J., Gilmer, J., Muelly, M., Goodfellow, I., Hardt, M., & Kim, B. (2018) · arXiv · doi:10.48550/arxiv.1810.03292 · Cited by 1,294 · cited 1×

    “Others have explored alternative ways of comparing the behavior of proposed explanation methods (Adebayo et al., 2018).”
    Background
  • Long Short-Term Memory-Networks for Machine Reading

    Cheng, J., Dong, L., & Lapata, M. (2016) · arXiv · doi:10.48550/arxiv.1601.06733 · Cited by 1,198 · cited 2×

    “, 2017), as well as cases where a single attention layer is responsible for producing more than one attended representation, such as in self-attention (Cheng et al., 2016).”
    Background
  • Long Short-Term Memory-Networks for Machine Reading

    Cheng, J., Dong, L., & Lapata, M. (2016) · Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d16-1053 · Cited by 1,172 · cited 2×

    “, 2017), as well as cases where a single attention layer is responsible for producing more than one attended representation, such as in self-attention (Cheng et al., 2016).”
    Background
  • Interpretation of Neural Networks is FragilePreprint

    Ghorbani, A., Abid, A., & Zou, J. Y. (2017) · AAAI Conference on Artificial Intelligence · doi:10.48550/arxiv.1710.10547 · Cited by 1,153 · cited 1×

    “In this work, we focus on whether the attention weights’ suggested importance ranking of I faithfully describes why the model produced its output, echoing existing work on explanation brittleness for other model components (Ghorbani et al., 2017).”
    Background
  • A Class of Submodular Functions for Document Summarization

    Lin, H. Z., & Bilmes, J. (2011) · Cited by 960

  • Rationalizing Neural Predictions

    Lei, T., Barzilay, R., & Jaakkola, T. (2016) · Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d16-1011 · Cited by 777 · cited 2×

    “…inputs necessary to uphold a decision is not new; Li et al. (2016) use reinforcement learning to attempt to construct such a minimal set of words, Lei et al. (2016) train an encoder to constrain the input prior to clas-sification, and much of the work that has been done on extractive…”
    Background
  • Interpretation of Neural Networks Is Fragile

    Ghorbani, A., Abid, A., & Zou, J. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v33i01.33013681 · Cited by 765

  • Effective Approaches to Attention-based Neural Machine Translation

    Luong, M.-T., Pham, H., & Manning, C. D. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1508.04025 · Cited by 655 · cited 1×

    “For many different problems — to name a few, machine translation (Luong et al., 2015), syntactic parsing (Vinyals et al., 2015), reading comprehension (Hermann et al., 2015), and language modeling (Liu and Lapata, 2018) — incorporating attention mechanisms into models has proven beneficial for…”
    Background
  • Visualizing and Understanding Neural Models in NLP

    Li, J., Chen, X., Hovy, E., & Jurafsky, D. (2016) · Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies · doi:10.18653/v1/n16-1082 · Cited by 625 · cited 1×

    “The second ranking scheme, inspired by Li et al. (2015) and Feng et al. (2018), is to order the attention weights by the gradient of the decision function with respect to each calculated attention weight, in descending order.”
    Background
  • Right for the Right Reasons: Training Differentiable Models by Constraining their Explanations

    Ross, A. S., Hughes, M. C., & Doshi-Velez, F. (2017) · Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2017/371 · Cited by 515 · cited 1×

    “Note, however, that we still do not take either as a gold-standard indicator of importance to the model, as with the gradient in Ross et al. (2017) and Melis and Jaakkola (2018), but merely as an alternative ordering method.”
    Method
  • Jointly modeling aspects, ratings and sentiments for movie recommendation (JMARS)

    Diao, Q., Qiu, M., Wu, C.-Y., Smola, A. J., Jiang, J., & Wang, C. (2014) · Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining · doi:10.1145/2623330.2623758 · Cited by 447 · cited 3×

    “We note that our IMDB data and Yelp data are different sets of reviews from those used by Yang et al. (2016), so our reported performances are not directly comparable to theirs.”
    BackgroundMethod
  • From Softmax to Sparsemax: A Sparse Model of Attention and Multi-Label Classification

    Martins, A. F. T., & Astudillo, R. F. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1602.02068 · Cited by 431 · cited 3×

    “Given attention layer ℓ ’s learned parameters, element 𝑖 of a sequence, and its encoded representation 𝐡 𝑖 , the attention weight 𝛼 𝑖 is computed using ℓ ’ s learned context vector 𝐜 ℓ as follows: 𝑗 4 Alternatives such as sparse attention (Martins and Astudillo, 2016) and unnormalized attention (Ji and Smith, 2017) ha…”
    BackgroundMethod
  • Understanding Neural Networks through Representation Erasure

    Li, J., Monroe, W., & Jurafsky, D. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1612.08220 · Cited by 412 · cited 2×

    “The idea of a minimal set of inputs necessary to uphold a decision is not new; Li et al. (2016) use reinforcement learning to attempt to construct such a minimal set of words, Lei et al. (2016) train an encoder to constrain the input prior to clas-sification, and much of the work that has been done…”
    BackgroundMethod
  • Towards Robust Interpretability with Self-Explaining Neural Networks

    Alvarez-Melis, D., & Jaakkola, T. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1806.07538 · Cited by 397 · cited 1×

    “(2017) and Melis and Jaakkola (2018), but merely as an alternative ordering method.”
    Background
  • Pathologies of Neural Models Make Interpretations Difficult

    Feng, S., Wallace, E., Grissom II, A., Iyyer, M., Rodriguez, P., & Boyd-Graber, J. (2018) · Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d18-1407 · Cited by 245 · cited 6×

    “To examine the importance of ℐ′ , we run the classification layer of the model twice (Figure 1): once without any modification, and once after renormalizing the attention distribution with ℐ′ ’s attention weights zeroed out, similar to other erasure-based work (Li et al., 2016; Feng et al., 2018).”
    BackgroundMethodInfluential
  • Comparing Automatic and Human Evaluation of Local Explanations for Text Classification

    Nguyen, D. (2018) · Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers) · doi:10.18653/v1/n18-1097 · Cited by 173 · cited 1×

    “Nguyen (2018) leveraged the idea of minimal sets in a much more similar way to our work, comparing different input importance orderings.”
    Result
  • Visualizing and Understanding Neural Machine Translation

    Ding, Y., Liu, Y., Luan, H., & Sun, M. (2017) · Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/p17-1106 · Cited by 168 · cited 1×

    “The existence of so much work on visualizing attention weights is a testament to attention’s popularity in this regard; to name just a few examples of these weights being examined to understand a model, recent work has focused on goals from explaining and debugging the current system’s decision (Lee et al., 2017; Ding…”
    Background
  • Visualizing and Understanding Neural Models in NLPPreprint

    Li, J., Chen, X., Hovy, E. H., et al. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1506.01066 · Cited by 165 · cited 1×

    “The second ranking scheme, inspired by Li et al. (2015) and Feng et al. (2018), is to order the attention weights by the gradient of the decision function with respect to each calculated attention weight, in descending order.”
    Background
  • Neural Discourse Structure for Text Categorization

    Ji, Y., & Smith, N. A. (2017) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/p17-1092 · Cited by 157

    Self-citation
  • Learning Structured Text Representations

    Liu, Y., & Lapata, M. (2018) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00005 · Cited by 145 · cited 1×

    “…many different problems—to name a few, machine translation (Luong et al., 2015), syntactic parsing (Vinyals et al., 2015), reading comprehension (Hermann et al., 2015), and language modeling (Liu and Lapata, 2018)—incorporating attention mechanisms into models has proven benefi-cial for performance.”
    Background
  • Before Name-Calling: Dynamics and Triggers of Ad Hominem Fallacies in Web Argumentation

    Habernal, I., Wachsmuth, H., Gurevych, I., & Stein, B. (2018) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.18653/v1/n18-1036 · Cited by 128 · cited 2×

    “, 2017) to distilling important traits of a dataset (Yang et al., 2017; Habernal et al., 2018).”
    Background
  • SPINE: SParse Interpretable Neural Embeddings

    Subramanian, A., Pruthi, D., Jhamtani, H., Berg-Kirkpatrick, T., & Hovy, E. H. (2018) · AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI)) · doi:10.1609/aaai.v32i1.11935 · Cited by 107

  • Explaining Predictions of Non-Linear Classifiers in NLP

    Arras, L., Horn, F., Montavon, G., Müller, K.-R., & Samek, W. (2016) · Proceedings of the 1st Workshop on Representation Learning for NLP · doi:10.18653/v1/w16-1601 · Cited by 91 · cited 1×

    “For example, recent work has focused on which training instances (Koh and Liang, 2017) or which human-interpretable features were most relevant for a particular decision (Ribeiro et al., 2016; Arras et al., 2016).”
    Background
  • Interpreting Recurrent and Attention-Based Neural Models: a Case Study on Natural Language Inference

    Ghaeini, R., Fern, X., & Tadepalli, P. (2018) · Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d18-1537 · Cited by 91 · cited 1×

    “Since attention calculates a distribution over inputs, prior work has used attention as a tool for interpretation of model decisions (Wang et al., 2016; Lee et al., 2017; Lin et al., 2017; Ghaeini et al., 2018).”
    Method
  • Satirical News Detection and Analysis using Attention Mechanism and Linguistic Features

    Yang, F., Mukherjee, A., & Dragut, E. C. (2017) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/d17-1211 · Cited by 83 · cited 2×

    “…name just a few examples of these weights being examined to understand a model, recent work has focused on goals from explaining and debugging the current system’s decision (Lee et al., 2017; Ding et al., 2017) to distilling important traits of a dataset (Yang et al., 2017; Habernal et al., 2018).”
    Background
  • Interactive Visualization and Manipulation of Attention-based Neural Machine Translation

    Lee, J., Shin, J.-H., & Kim, J.-S. (2017) · Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing: System Demonstrations · doi:10.18653/v1/d17-2021 · Cited by 67 · cited 3×

    “The existence of so much work on visualizing attention weights is a testament to attention’s popularity in this regard; to name just a few examples of these weights being examined to understand a model, recent work has focused on goals from explaining and debugging the current system’s decision (Lee et al., 2017; Ding…”
    BackgroundMethod
  • A Compositional and Interpretable Semantic Space

    Fyshe, A., Wehbe, L., Talukdar, P. P., Murphy, B., & Mitchell, T. M. (2015) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.3115/v1/n15-1004 · Cited by 63 · cited 1×

    “Yet another line of work focuses on aligning models with human feedback for what is interpretable (Fyshe et al., 2015; Subramanian et al., 2017), which could refine our idea of what defines a high-quality explanation derived from attention.”
    Background
  • Interpreting Recurrent and Attention-Based Neural Models: a Case Study on Natural Language InferencePreprint

    Ghaeini, R., Fern, X. Z., & Tadepalli, P. V. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1808.03894 · Cited by 10 · cited 1×

    “Since attention calculates a distribution over inputs, prior work has used attention as a tool for interpretation of model decisions (Wang et al., 2016; Lee et al., 2017; Lin et al., 2017; Ghaeini et al., 2018).”
    Method
  • Satirical News Detection and Analysis using Attention Mechanism and Linguistic FeaturesPreprint

    Yang, F., Mukherjee, A., & Dragut, E. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1709.01189 · Cited by 5 · cited 2×

    “…name just a few examples of these weights being examined to understand a model, recent work has focused on goals from explaining and debugging the current system’s decision (Lee et al., 2017; Ding et al., 2017) to distilling important traits of a dataset (Yang et al., 2017; Habernal et al., 2018).”
    Background

Research works printed but not linked: 4

  1. Philipp Koehn and Rebecca Knowles. 2017. Six chal- lenges for neural machine translation. In Pro- ceedings of the First Workshop on Neural Machine Translation.
  2. Zachary C Lipton. 2016. The mythos of model inter- pretability. arXiv preprint arXiv:1606.03490. · arXiv
  3. Yang Liu and Mirella Lapata. 2018. Learning struc- tured text representations. Transactions of the Asso- ciation of Computational Linguistics, 6:63-75.
  4. Oriol Vinyals, Łukasz Kaiser, Terry Koo, Slav Petrov, Ilya Sutskever, and Geoffrey Hinton. 2015. Gram- mar as a foreign language. In Advances in Neural Information Processing Systems.

Cited by

See all 697

Metadata

Authors
Serrano, S., & Smith, N. A.
Affiliations
University of Washington (US)
Venue
Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics
Published
2019
Pages
2931-2951
ISSN
0736-587X (ISSN-L)
OpenAlex
W2948140294
Cited by
690
References
39
Citations by year
2026: 63 · 2025: 85 · 2024: 82 · 2023: 102 · 2022: 119 · 2021: 143 · 2020: 80 · 2019: 15
FWCI
34.15
Type
conference paper
Language
en
Open access
Licence
cc-by
Locations
  • Proceedings of the conference - Association for Computational Linguistics. Meeting (publishedVersion, cc-by)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
  • Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
  • Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
Keywords
attention mechanism, interpretability, text classification, model interpretability, feature importance, explainable AI

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.