Published on Mon Sep 30 2019

Automatic Fact-guided Sentence Modification

Darsh J Shah, Tal Schuster, Regina Barzilay

Online encyclopediae like Wikipedia contain large amounts of text that need frequent corrections and updates. In this paper, we focus on rewriting such dynamically changing articles. The output must be consistent with the new information and fit into the rest of the document.

3
0
0
Abstract

Online encyclopediae like Wikipedia contain large amounts of text that need frequent corrections and updates. The new information may contradict existing content in encyclopediae. In this paper, we focus on rewriting such dynamically changing articles. This is a challenging constrained generation task, as the output must be consistent with the new information and fit into the rest of the existing document. To this end, we propose a two-step solution: (1) We identify and remove the contradicting components in a target text for a given claim, using a neutralizing stance model; (2) We expand the remaining text to be consistent with the given claim, using a novel two-encoder sequence-to-sequence model with copy attention. Applied to a Wikipedia fact update dataset, our method successfully generates updated sentences for new claims, achieving the highest SARI score. Furthermore, we demonstrate that generating synthetic data through such rewritten sentences can successfully augment the FEVER fact-checking training dataset, leading to a relative error reduction of 13%.

Thu Jul 02 2020
NLP
Fact-based Text Editing
The goal is to revise a given document to better describe the facts in a knowledge base. A straightforward approach to address the problem would be to employ an encoder-decoder model. We propose a new neural network architecture for fact-based text editing.
0
0
0
Wed Jun 02 2021
NLP
Evidence-based Factual Error Correction
This paper introduces the task of factual error correction: performing edits to a claim so that the generated rewrite is better supported by evidence. We achieve this by employing a two-stage distant supervision approach that incorporates evidence into masked claims when generating corrections.
12
0
0
Thu Dec 31 2020
NLP
Evidence-based Factual Error Correction
This paper introduces the task of factual error correction: performing edits to a claim so that the generated rewrite is better supported by evidence. We achieve this by employing a two-stage distant supervision approach that incorporates evidence into masked claims when generating corrections.
3
0
0
Mon Mar 15 2021
NLP
Get Your Vitamin C! Robust Fact Verification with Contrastive Evidence
VitaminC is a benchmark infused with challenging cases that require fact verification models to discern and adjust to slight factual changes. Unlike previous resources, the examples in VitaminC are nearly identical in language and content, with the exception that one supports a given claim while the other does not.
1
56
200
Sat Sep 14 2019
NLP
ALTER: Auxiliary Text Rewriting Tool for Natural Language Generation
ALTER is an auxiliary text rewriting tool. It can be used for natural language generation tasks. It is characterized by two features: recording of word-level revision histories and flexible auxiliary edit support.
0
0
0
Fri Apr 16 2021
NLP
Editing Factual Knowledge in Language Models
Some facts can be mistakenly induced or become obsolete over time. We present a method that can be used to edit this knowledge and fix 'bugs' or 'unexpected predictions' This method does not need expensive re-training or fine-tuning.
1
33
216
Thu Oct 11 2018
NLP
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
BERT is designed to pre-train deep                bidirectional representations from unlabeled text. It can be fine-tuned with just one additional output layer to create state-of-the-art models for a wide range of tasks.
13
8
15
Tue Aug 28 2018
NLP
WikiAtomicEdits: A Multilingual Corpus of Wikipedia Edits for Modeling Language and Discourse
We release a corpus of 43 million atomic edits across 8 languages. These are mined from Wikipedia edit history and consist of instances in which a human editor has inserted a single contiguous phrase into, or deleted a single contiguous phrase from, an existing sentence.
0
0
0
Wed Aug 11 2010
NLP
For the sake of simplicity: Unsupervised extraction of lexical simplifications from Wikipedia
We report on work in progress on extracting lexical simplifications. We consider two main approaches: (1)iving simplification probabilities via an edit model that accounts for a mixture of different operations, and (2) using metadata to focus on edits that are more likely to be simplification operations.
0
0
0
Wed May 10 2017
NLP
A Minimal Span-Based Neural Constituency Parser
We present a minimal neural model for constituency parsing based on independent scoring of labels and spans. We show that this model is compatible with classical dynamic programming techniques, but alsomits a novel greedy top-down inference algorithm. We demonstrate that both prediction schemes are competitive with recent work.
0
0
0
Wed Aug 14 2019
NLP
Towards Debiasing Fact Verification Models
Claim-only classifiers perform competitively with top evidence-aware models.
0
0
0
Tue Apr 17 2018
NLP
Delete, Retrieve, Generate: A Simple Approach to Sentiment and Style Transfer
We consider the task of text attribute transfer: transforming a sentence to alter a specific attribute. Previous adversarial methods have struggled to produce high-quality outputs. Our strongest method extracts content words by deleting phrases associated with the original attribute.
0
0
0
Mon Mar 15 2021
NLP
Get Your Vitamin C! Robust Fact Verification with Contrastive Evidence
VitaminC is a benchmark infused with challenging cases that require fact verification models to discern and adjust to slight factual changes. Unlike previous resources, the examples in VitaminC are nearly identical in language and content, with the exception that one supports a given claim while the other does not.
1
56
200
Mon Mar 22 2021
NLP
Nutri-bullets: Summarizing Health Studies by Composing Segments
0
0
0
Fri Jul 31 2020
Artificial Intelligence
Neural Language Generation: Formulation, Methods, and Evaluation
Recent advances in neural network-based generative modeling have reignited the hopes in having computer systems capable of seamlessly conversing with humans. While the field of natural language generation is evolving rapidly, there are still many open challenges to address.
0
0
0
Sun Apr 18 2021
NLP
Generating Related Work
0
0
0
Mon Apr 19 2021
NLP
\textit{NewsEdits}: A Dataset of Revision Histories for News Articles (Technical Report: Data Processing)
This is the first publicly available dataset of news article revision histories. It contains 1,278,804 articles with 4,609,430versions from over 22 English- and French-language newspaper sources. Across version pairs, we count 10.9 million added sentences and 6.8 million removed
1
0
0
Thu Dec 31 2020
NLP
Evidence-based Factual Error Correction
This paper introduces the task of factual error correction: performing edits to a claim so that the generated rewrite is better supported by evidence. We achieve this by employing a two-stage distant supervision approach that incorporates evidence into masked claims when generating corrections.
3
0
0