paper-with-me

Papers

On Exposure Bias, Hallucination and Domain Shift in Neural Machine Translation

2020-05-07 · ACL 2020 6 · Chaojun Wang, Rico Sennrich

The standard training algorithm in neural machine translation (NMT) suffers from exposure bias, and alternative algorithms have been proposed to mitigate this. However, the practical impact of exposure bias is under debate. In this paper, we link exposure bias to another well-known problem in NMT, namely the tendency to generate hallucinations under domain shift. In experiments on three datasets with multiple test domains, we show that exposure bias is partially to blame for hallucinations, and that training with Minimum Risk Training, which avoids exposure bias, can mitigate this. Our analysis explains why exposure bias is more problematic under domain shift, and also links exposure bias to the beam search problem, i.e. performance deterioration with increasing beam size. Our results provide a new justification for methods that reduce exposure bias: even if they do not increase performance on in-domain test sets, they can increase model robustness to domain shift.

📄 PDF Abstract BibTeX arXiv:2005.03642

Code (1)

zippotju/Exposure-Bias-Hallucination-Domain-Shift 공식 구현

Tasks

HallucinationMachine TranslationNMTTranslation

Similar Papers 제목 키워드 기반

BRIDO: Bringing Democratic Order to Abstractive Summarization

2025-02-25 · Junhyun Lee, Harshith Goka, Hyeonmok Ko

Hallucination refers to the inaccurate, irrelevant, and inconsistent text generated from large language models (LLMs). While the LLMs have shown great promise in a variety of tasks, the issue of hallucination still remai…

Abstractive Text SummarizationContrastive LearningHallucinationText Summarization

The Highs and Lows of Simple Lexical Domain Adaptation Approaches for Neural Machine Translation

2021-01-02 · EMNLP (insights) 2021 11 · Nikolay Bogoychev, Pinzhen Chen

Machine translation systems are vulnerable to domain mismatch, especially in a low-resource scenario. Out-of-domain translations are often of poor quality and prone to hallucinations, due to exposure bias and the decoder…

DecoderDomain AdaptationLanguage ModelingLanguage Modelling+3

Why Exposure Bias Matters: An Imitation Learning Perspective of Error Accumulation in Language Generation

2022-04-03 · Findings (ACL) 2022 5 · Kushal Arora, Layla El Asri, Hareesh Bahuleyan, Jackie Chi Kit Cheung

Current language generation models suffer from issues such as repetition, incoherence, and hallucinations. An often-repeated hypothesis is that this brittleness of generation models is caused by the training and the gene…

Imitation LearningText Generation

Progressive Translation: Improving Domain Robustness of Neural Machine Translation with Intermediate Sequences

2023-05-16 · Chaojun Wang, Yang Liu, Wai Lam

Previous studies show that intermediate supervision signals benefit various Natural Language Processing tasks. However, it is not clear whether there exist intermediate signals that benefit Neural Machine Translation (NM…

Inductive BiasMachine TranslationMulti-Task LearningNMT+1

Alleviating Exposure Bias in Diffusion Models through Sampling with Shifted Time Steps

2023-05-24 · Mingxiao Li, Tingyu Qu, Ruicong Yao, Wei Sun 외

Diffusion Probabilistic Models (DPM) have shown remarkable efficacy in the synthesis of high-quality images. However, their inference process characteristically requires numerous, potentially hundreds, of iterative steps…

Denoising