LS-Tree: Model Interpretation When the Data Are Linguistic

We study the problem of interpreting trained classification models in the setting of linguistic data sets. Leveraging a parse tree, we propose to assign least-squares-based importance scores to each word of an instance by exploiting syntactic constituency structure. We establish an axiomatic characterization of these importance scores by relating them to the Banzhaf value in coalitional game theory. Based on these importance scores, we develop a principled method for detecting and quantifying interactions between words in a sentence. We demonstrate that the proposed method can aid in interpretability and diagnostics for several widely-used language models.

[1]  S. Larson.  The shrinkage of the coefficient of multiple correlation. , 1931 .

[2]  J. Sherman,et al.  Adjustment of an Inverse Matrix Corresponding to a Change in One Element of a Given Matrix , 1950 .

[3]  Pradeep Dubey,et al.  Mathematical Properties of the Banzhaf Power Index , 1979, Math. Oper. Res..

[4]  L. Shapley.  A Value for n-person Games , 1988 .

[5]  Peter L. Hammer,et al.  Approximations of pseudo-Boolean functions; applications to game theory , 1992, ZOR Methods Model. Oper. Res..

[6]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[7]  On an axiomatization of the Banzhaf value without the additivity axiom , 1997 .

[8]  R. Dennis Cook,et al.  Detection of Influential Observation in Linear Regression , 2000, Technometrics.

[9]  Alon Lavie,et al.  A Classifier-Based Parser with Linear Run-Time Complexity , 2005, IWPT.

[10]  Stephen Clark,et al.  Transition-Based Parsing of the Chinese Treebank using a Global Discriminative Model , 2009, IWPT.

[11]  Motoaki Kawanabe,et al.  How to Explain Individual Classification Decisions , 2009, J. Mach. Learn. Res..

[12]  Erik Strumbelj,et al.  An Efficient Explanation of Individual Classifications using Game Theory , 2010, J. Mach. Learn. Res..

[13]  Christopher Potts,et al.  Learning Word Vectors for Sentiment Analysis , 2011, ACL.

[14]  Joakim Nivre,et al.  A Dynamic Oracle for Arc-Eager Dependency Parsing , 2012, COLING.

[15]  Yue Zhang,et al.  Fast and Accurate Shift-Reduce Constituent Parsing , 2013, ACL.

[16]  Christopher Potts,et al.  Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank , 2013, EMNLP.

[17]  I. Katsev.  The Least Square Values for Games with Restricted Cooperation , 2013 .

[18]  Yoon Kim,et al.  Convolutional Neural Networks for Sentence Classification , 2014, EMNLP.

[19]  Andrew Zisserman,et al.  Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps , 2013, ICLR.

[20]  Jeffrey Pennington,et al.  GloVe: Global Vectors for Word Representation , 2014, EMNLP.

[21]  Alexander Binder,et al.  On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation , 2015, PloS one.

[22]  Fei-Fei Li,et al.  Visualizing and Understanding Recurrent Networks , 2015, ArXiv.

[23]  Yoshua Bengio,et al.  Show, Attend and Tell: Neural Image Caption Generation with Visual Attention , 2015, ICML.

[24]  Xiang Zhang,et al.  Character-level Convolutional Networks for Text Classification , 2015, NIPS.

[25]  Koray Kavukcuoglu,et al.  Multiple Object Recognition with Visual Attention , 2014, ICLR.

[26]  Alex Graves,et al.  DRAW: A Recurrent Neural Network For Image Generation , 2015, ICML.

[27]  Wei Xu,et al.  ABC-CNN: An Attention Based Convolutional Neural Network for Visual Question Answering , 2015, ArXiv.

[28]  Kate Saenko,et al.  Ask, Attend and Answer: Exploring Question-Guided Spatial Attention for Visual Question Answering , 2015, ECCV.

[29]  Alexander J. Smola,et al.  Stacked Attention Networks for Image Question Answering , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[30]  Regina Barzilay,et al.  Rationalizing Neural Predictions , 2016, EMNLP.

[31]  Daniel Jurafsky,et al.  Understanding Neural Networks through Representation Erasure , 2016, ArXiv.

[32]  Carlos Guestrin,et al.  "Why Should I Trust You?": Explaining the Predictions of Any Classifier , 2016, ArXiv.

[33]  Yair Zick,et al.  Algorithmic Transparency via Quantitative Input Influence: Theory and Experiments with Learning Systems , 2016, 2016 IEEE Symposium on Security and Privacy (SP).

[34]  Xinlei Chen,et al.  Visualizing and Understanding Neural Models in NLP , 2015, NAACL.

[35]  Avanti Shrikumar,et al.  Learning Important Features Through Propagating Activation Differences , 2017, ICML.

[36]  Lukasz Kaiser,et al.  Attention is All you Need , 2017, NIPS.

[37]  Scott Lundberg,et al.  A Unified Approach to Interpreting Model Predictions , 2017, NIPS.

[38]  Ankur Taly,et al.  Axiomatic Attribution for Deep Networks , 2017, ICML.

[39]  Wesley De Neve,et al.  Explaining Character-Aware Neural Networks for Word-Level Prediction: Do They Discover Linguistic Rules? , 2018, EMNLP.

[40]  Le Song,et al.  L-Shapley and C-Shapley: Efficient Model Interpretation for Structured Data , 2018, ICLR.

[41]  Ming-Wei Chang,et al.  BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding , 2019, NAACL.