| Challenge: | Attention mechanisms have seen wide adoption in neural NLP models. |
| Approach: | They perform extensive experiments to assess the degree to which attention weights provide meaningful "explanations" they find that attention weighted inputs are often uncorrelated with gradient-based measures of feature importance . |
| Outcome: | The proposed model is based on a distribution over attended-to input units . the findings show that attention weights are often uncorrelated with features . |
Similar Papers
Is Attention Interpretable? (P19-1)
Copied to clipboard
| Challenge: | Attention mechanisms have recently boosted performance on a range of NLP tasks. |
| Approach: | They propose to manipulate attention weights in text classification models and analyze the resulting differences in their predictions. |
| Outcome: | The proposed approach improves models' predictions by using gradient-based rankings of attention weights. |
Attention is not not Explanation (D19-1)
Copied to clipboard
| Challenge: | Attention mechanisms play a central role in NLP systems, especially within recurrent neural network (RNN) models. |
| Approach: | They propose to use a simple uniform-weights baseline, a variance calibration and a diagnostic framework to determine when/whether attention can be used as explanation in RNN models. |
| Outcome: | The proposed tests show that even reliable adversarial distributions don't perform well on the simple diagnostic, indicating that prior work does not disprove the usefulness of attention mechanisms for explainability. |
Is Attention Explanation? An Introduction to the Debate (2022.acl-long)
Copied to clipboard
Adrien Bibal, Rémi Cardon, David Alfter, Rodrigo Wilkens, Xiaoou Wang, Thomas François, Patrick Watrin
| Challenge: | Attention has been used in various tasks of NLP and other fields of machine learning to increase performance and provide some explanations. |
| Approach: | They propose to use attention as an explanation for deep learning models to increase performance . they propose to apply attention weights to queries and queries based on scalar scores . |
| Outcome: | The proposed model can be used to increase performance while providing some explanations. |
Learning to Deceive with Attention-Based Explanations (2020.acl-main)
Copied to clipboard
| Challenge: | Attention mechanisms are ubiquitous components in neural network architectures and are often claimed to confer interpretability. |
| Approach: | They propose a method for training models to produce deceptive attention masks by combining weights assigned to designated impermissible tokens with a weighted sum. |
| Outcome: | The proposed method reduces the weight assigned to designated impermissible tokens while still using them across multiple models and tasks. |
Why Attention is Not Explanation: Surgical Intervention and Causal Reasoning about Neural Models (2020.lrec-1)
Copied to clipboard
| Challenge: | a recent study finds brittleness in explanations obtained through attention mechanisms . a philosophy of science theory allows robust yet non-causal reasoning in explanation . |
| Approach: | They propose to use philosophy of science to examine the state-of-the-art in explanation for NLP models . they argue that it is impossible to explain attention-based learning by attention mechanisms . |
| Outcome: | The proposed model selection criteria are based on philosophy of science theories . the proposed model is based upon a model that is more explainable than a classical model . |
Towards Transparent and Explainable Attention Models (2020.acl-main)
Copied to clipboard
Akash Kumar Mohankumar, Preksha Nema, Sharan Narasimhan, Mitesh M. Khapra, Balaji Vasan Srinivasan, Balaraman Ravindran
| Challenge: | Recent studies on interpretability of attention distributions have led to notions of faithful and plausible explanations for a model’s predictions. |
| Approach: | They propose to modify LSTM cells to ensure that the hidden representations learned at different time steps are diverse. |
| Outcome: | The proposed model can provide a faithful explanation if a higher attention weight implies a greater impact on the model’s prediction. |
How does Attention Affect the Model? (2021.findings-acl)
Copied to clipboard
| Challenge: | Existing studies on the effectiveness of attention in NLP do not consider changes in semantic capability of different components. |
| Approach: | They propose a framework that exploits a convex hull representation of sequence semantics in an n-dimensional Semantic Euclidean Space and defines indicators to capture the impact of attention on sequence semantic. |
| Outcome: | The proposed framework exploits a convex hull representation of sequence semantics in an n-dimensional Semantic Euclidean Space and defines indicators to capture the impact of attention on sequence semantic. |
Is Sparse Attention more Interpretable? (2021.acl-short)
Copied to clipboard
| Challenge: | Sparse attention has been claimed to increase model interpretability . however, the attention distribution is typically over representations internal to the model rather than the inputs themselves . |
| Approach: | They conduct experiments to understand how sparsity affects our ability to use attention as an explainability tool. |
| Outcome: | The proposed model does not map to a sparse set of influential inputs, but rather to fewer inputs. |
Understanding Attention for Text Classification (2020.acl-main)
Copied to clipboard
| Challenge: | Existing studies have focused on whether local attention weights reflect the importance of input representations. |
| Approach: | They propose to analyze for each word token the following two quantities: its polarity score and its attention score, where the latter is a global assessment on the token’s significance. |
| Outcome: | The proposed model can be improved under conditions where the interplay between the two quantities can contribute towards model performance. |
Interrogating the Explanatory Power of Attention in Neural Machine Translation (D19-56)
Copied to clipboard
| Challenge: | Attention models are often used to justify the model’s decision in generating a token but it has not been rigorously established to what extent attention is a reliable source of information in NMT. |
| Approach: | They propose to use attention models to modify crucial aspects of the trained attention model to produce function and content words in the translation process. |
| Outcome: | The proposed models preserve function and content words in the translation process compared to state-of-the-art models. |