Skip to content
conftrace_
2018 EMNLP EMNLP 2018

Learning Explanations from Language Data

Abstract

AbstractPatternAttribution is a recent method, introduced in the vision domain, that explains classifications of deep neural networks. We demonstrate that it also generates meaningful interpretations in the language domain.

📈 Trend Setter - Interpretability
🐣 Hot Topic Early Bird - model interpretability

Authors

David Harbecke, Robert Schwarzenberg, Christoph Alt

Topics

Artificial Intelligence > Core AI > Interpretability Deep Learning > Architectures > Neural Networks Deep Learning > Optimization & Theory > Interpretability

Keywords

natural language processing feature attribution model interpretability deep neural network neural network pattern attribution
Download PDF

More from EMNLP 2018

All EMNLP 2018 papers →
An Analysis of Encoder Representations in Transformer-Based Machine Translation
Evaluating Grammaticality in Seq2seq Models with a Broad Coverage HPSG Grammar: A Case Study on Machine Translation
Context-Free Transductions with Neural Stacks
How much should you ask? On the question structure in QA systems.
Does it care what you asked? Understanding Importance of Verbs in Deep Learning QA System
Interpretable Textual Neuron Representations for NLP