Adversarial Decomposition of Text Representation

Alexey Romanov, Anna Rumshisky, Anna Rogers, David Donahue
8/27/2018
cs.CL

Official code: text-machine-lab/adversarial_decomposition (29 stars, last commit 2022-12-08).

Abstract

In this paper, we present a method for adversarial decomposition of text representation. This method can be used to decompose a representation of an input sentence into several independent vectors, each of them responsible for a specific aspect of the input sentence. We evaluate the proposed method on two case studies: the conversion between different social registers and diachronic language change. We show that the proposed method is capable of fine-grained controlled change of these aspects of the input sentence. It is also learning a continuous (rather than categorical) representation of the style of the sentence, which is more linguistically realistic. The model uses adversarial-motivational training and includes a special motivational loss, which acts opposite to the discriminator and encourages a better decomposition. Furthermore, we evaluate the obtained meaning embeddings on a downstream task of paraphrase detection and show that they significantly outperform the embeddings of a regular autoencoder.

View on arXivView PDF

Code Implementations(1)

text-machine-lab/adversarial_decompositionOfficial100%

The code for the paper "Adversarial Decomposition of Text Representation", NAACL 2019

296Python, Jupyter NotebookApr 3, 20193 years ago

Similar Papers

A Constraint-Enforcing Reward for Adversarial Attacks on Text Classifiers

82% match2024

Interpretable Adversarial Perturbation in Input Embedding Space for Text

82% match1 repo422018

Adversarial Example Generation with Syntactically Controlled Paraphrase Networks

82% match1 repo1682018

Don't Retrain, Just Rewrite: Countering Adversarial Perturbations by Rewriting Text

81% match1 repo2023

Adversarial Training for Aspect-Based Sentiment Analysis with BERT

80% match1 repo402020

Certified Robustness to Adversarial Word Substitutions

80% match2 repos382019

Cite this paper

@article{romanov2018adversarial,
  title  = {Adversarial Decomposition of Text Representation},
  author = {Alexey Romanov and Anna Rumshisky and Anna Rogers and David Donahue},
  year   = {2018},
  eprint = {1808.09042},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1808.09042}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1808.09042.svg)](https://codeofpaper.com/paper/1808.09042)

Discussion