ProtoTEx: Explaining Model Decisions with Prototype Tensors

PROTOTEX: Explaining Model Decisions with Prototype Tensors Anubrata Das1∗ Chitrank Gupta2 ∗ Venelin Kovatchev1 Matthew Lease1 Junyi Jessy Li3 1School of Information and 3Dept. of Linguistics, The University of Texas at Austin 2Dept. of Computer Science, Indian Institute of Technology Bombay {anubrata, venelin, ml, jessy}@utexas.edu, chigupta2011@gmail.com Abstract We present PROTOTEX, a novel white-box NLP classification architecture based on pro- totype networks (Li et al., 2018). PROTOTEX faithfully explains model decisions based on prototype tensors that encode latent clusters of training examples. At inference time, classi- fication decisions are based on the distances between the input text and the prototype ten- sors, explained via the training examples most similar to the most influential prototypes. We also describe a novel interleaved training algo- rithm that effectively handles classes character- ized by the absence of indicative features. On a propaganda detection task, PROTOTEX accu- racy matches BART-large and exceeds BERT- large with the added benefit of providing faith- ful explanations. A user study also shows that prototype-based explanations help non-experts to better recognize propaganda in online news. 1 Introduction Neural models for NLP have yielded significant gains in predictive accuracy across a wide range of tasks. However, these state-of-the-art models are typically less interpretable than simpler, traditional models, such as decision trees or nearest-neighbor approaches. In general, less interpretable models can be more difficult for people to use, trust, and adopt in practice. Consequently, there is growing interest in going beyond simple “black-box” model accuracy to instead design models that are both highly accurate and human-interpretable. While much research on white-box explainable models focuses on attributing parts of the input (e.g., word sequences) to a model’s prediction (Xu et al., 2015; Lei et al., 2016; Bastings et al., 2019; Jain et al., 2020; Glockner et al., 2020), there is much debate around their faithfulness and reliabil- ity (Serrano and Smith, 2019; Jain and Wallace, 2019; Wiegreffe and Pinter, 2019; Pruthi et al., ∗Both authors contributed equally. 2020). Additionally, while such local explanations (if faithful) can be extremely useful in more in- tuitive tasks such as sentiment classification, that may not be the case for difficult tasks where human judgments may require a high degree of training or domain expertise. In such cases, understanding how models make their decisions for a particular input based on its training data can be insightful especially for engaging with users to develop an intuition on the model’s decision making process. In this paper, we propose Prototype Tensor Explainability Network (PROTOTEX)1 to faith- fully explain classification decisions in the tradi- tion of case-based reasoning (Kolodner, 1992). Our novel white-box NLP architecture augments proto- type classification networks (Li et al., 2018) with large-scale pretrained transformer language mod- els. Through a novel training regime, the network learns a set of prototype tensors that encode latent clusters of training examples. At inference time, classification decisions are entirely based on simi- larity to prototypes. This enables model predictions to be faithfully explained based on these prototypes, directly via similar training examples (i.e., those …

Content truncated to 500 words