Evaluating Factuality in Generation with Dependency-level Entailment

By Tanya Goyal, Greg Durrett. 2020

173 citations recorded by the source corpus.

Despite significant progress in text generation models, a serious limitation is their tendency to produce text that is factually inconsistent with information in the input. Recent work has studied whether textual entailment systems can be used to identify factual errors; however, these sentence-level entailment models are trained to solve a different problem than generation filtering and they do not localize which part of a generation is non-factual. In this paper, we propose a new formulation of entailment that decomposes it at the level of dependency arcs. Rather than focusing on aggregate decisions, we instead ask whether the semantic relationship manifested by individual dependency arcs in the generated output is supported by the input. Human judgments on this task are difficult to obtain; we therefore propose a method to automatically create data based on existing entailment or paraphrase corpora. Experiments show that our dependency arc entailment model trained on this data can identify factual inconsistencies in paraphrasing and summarization better than sentence-level methods or those based on question generation, while additionally localizing the erroneous parts of the generation.

Subfields: Artificial Intelligence, Natural Language Processing, Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

In Evaluation & Benchmarks

Citations in the corpus

Language Generation Models Can Cause Harm: So What Can We Do About It? An Actionable Survey (2022, methodology): “, 2022); Factual error detection (Kryscinski et al., 2020; Goyal and Durrett, 2020; Pagnoni et al., 2021) Disinformation Machine-generated text detection (Dugan et al.”
Survey of Hallucination in Natural Language Generation (2022, background): “Apart from generalizability, Goyal and Durrett [56] point out the potential limitation of using sentence-level entailment models – their incapability to localize which parts of the generation are erroneous.”
Length-Controlled AlpacaEval: A Simple Way to Debias Automatic Evaluators (2024, methodology): “Reference-free evaluation metrics Reference-free evaluation metrics have a long history, including classic methods (Louis & Nenkova, 2013) and more recent neural supervised learning methods (Kryscinski et al., 2020; Sinha et al., 2020; Goyal & Durrett, 2020).”
Understanding Factuality in Abstractive Summarization with FRANK: A Benchmark for Factuality Metrics (2021, background): “DAE (Goyal and Durrett, 2020): DAE is an entailment classification metric that operates on dependencies.”
CLIFF: Contrastive Learning for Improving Faithfulness and Factuality in Abstractive Summarization (2021, background): “To cover more diverse extrinsic and intrinsic errors (Goyal and Durrett, 2020), we extend M ASK E NT to contain relations ( M ASK R EL ).”
News Summarization and Evaluation in the Era of GPT-3 (2022, background): “, 2020) and trained factuality models (Kryscinski et al., 2020; Goyal and Durrett, 2020), similarly fail to adapt to this shift from the fine-tuned to zero-shot summarization space and need to be re-visited.”
FactGraph: Evaluating Factuality in Summarization with Semantic Graph Representations (2022, unknown): “Another line of work explores dependency arc entailment to improve the localization of subsentence-level errors within generated summaries (Goyal and Durrett, 2020).”
SummaC: Re-Visiting NLI-based Models for Inconsistency Detection in Summarization (2021, methodology): “More specifically, some of the model’s complexity scales with the number of sub-word units in the document (MNLI-doc, FactCC-CLS), some with the number of entities in a document (NER-Overlap, DAE, QuestEval), and some with number of sentences (the SUMMAC models).”
Are Factuality Checkers Reliable? Adversarial Meta-evaluation of Factuality in Summarization (2021, background): “, ROUGE) are not correlated well with factual consistency (Maynez et al., 2020; Goyal and Durrett, 2020).”
Correcting Diverse Factual Errors in Abstractive Summarization via Post-Editing and Language Model Infilling (2022, methodology): “Motivated by prior work (Goyal and Durrett, 2020) using lower ranked beam search candidates as a source for adversarial data, we use the lower ranked candidates to construct erroneous summaries from reference summaries.”
Annotating and Modeling Fine-grained Factuality in Summarization (2021, methodology): “The Dependency Arc Entailment (DAE) model (Goyal and Durrett, 2020a) evaluates factuality at the dependency arc level.”
Efficient Streaming Language Models with Attention Sinks (2023, background): “…language processing applications such as dialog systems (Schulman et al., 2022; Taori et al., 2023; Chiang et al., 2023), document summarization (Goyal & Durrett, 2020; Zhang et al., 2023a), code completion (Chen et al., 2021; Rozière et al., 2023) and question answering (Kamalloo et al., 2023).”
Benchmarking Large Language Models for News Summarization (2023, methodology): “Recent work (Goyal and Durrett, 2020) shows that the instruct-tuned GPT-3 Davinci model is better than ﬁnetuned LMs, but do not show the design decision that contribute to the improved performance.”
Hallucinated but Factual! Inspecting the Factuality of Hallucinations in Abstractive Summarization (2021, background): “(Goyal and Durrett, 2020) measure factual consistency by checking whether the semantic relationship manifested by individual dependency arcs in the generated summary is supported by the source document.”
Spurious Correlations in Reference-Free Evaluation of Text Generation (2022, methodology): “…of these reference-free evaluations achieve remarkably high correlations with human evaluations, raising hopes that they may soon become a viable alternative to expensive human evaluations (Kryscinski et al., 2020; Goyal and Durrett, 2020; Sinha et al., 2020; Phy et al., 2020; Gao et al., 2020).”
Improving Faithfulness in Abstractive Summarization with Contrast Candidate Generation and Selection (2021, background): “Recent studies explore categorical, contentbased analysis for measuring the faithfulness of summaries (Goyal and Durrett, 2020; Deutsch and Roth, 2020).”
A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions (2023, unknown): “…have proposed methods such as fine-tuning on adversarial datasets (Barrantes et al., 2020), decomposing the entailment decisions at the dependency arc level (Goyal and Durrett, 2020), and segmenting documents into sentence units then aggregating scores between sentence pairs (Laban et al., 2022).”
Evaluating Factuality in Text Simplification (2022, methodology): “, 2020b), and (2) DAE, which is another BERT-based model that classifies each dependency arc in the model output as entailing the source text or not (Goyal and Durrett, 2020).”
QAFactEval: Improved QA-Based Factual Consistency Evaluation for Summarization (2021, methodology): “Other work on entailment-based metrics has examined input granularity (Goyal and Durrett, 2020), trained on adversarial datasets (Barrantes et al., 2020), and explored entailment-based models as the backbone of others metrics such as BERTScore (Zhang et al., 2020b) as in Koto et al. (2021).”