mlsanityData to 14 Oct 2026

Revising with a Backward Glance: Regressions and Skips during Reading as Cognitive Signals for Revision Policies in Incremental Processing

arXiv 2310.18229 ↗First posted 27 Oct 2023cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

In NLP, incremental processors produce output in instalments, based on incoming prefixes of the linguistic input. Some tokens trigger revisions, causing edits to the output hypothesis, but little is known about why models revise when they revise. A policy that detects the time steps where revisions should happen can improve efficiency. Still, retrieving a suitable signal to train a revision policy is an open problem, since it is not naturally available in datasets. In this work, we investigate the appropriateness of regressions and skips in human reading eye-tracking data as signals to inform revision policies in incremental sequence labelling. Using generalised mixed-effects models, we find that the probability of regressions and skips by humans can potentially serve as useful predictors for revisions in BiLSTMs and Transformer models, with consistent results for various languages.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 18 of its 57 references that are in the map. First posted on arXiv 27 Oct 2023.

Builds on

18 of 57 references.

References · 18

  1. 2017Attention Is All You Needcited by 60,361
  2. 2020Stanza: A Python Natural Language Processing Toolkit for Many Human Languagescited by 723
  3. 2020On the Predictive Power of Neural Language Models for Human Real-Time Comprehension Behaviorcited by 119
  4. 2016Can neural machine translation do simultaneous translation?cited by 86
  5. 2020Re-translation versus Streaming for Simultaneous Translationcited by 31
  6. 2021Multilingual Language Models Predict Human Reading Behaviorcited by 29
  7. 2019Entity Recognition at First Sight: Improving NER with Eye Movement Informationcited by 26
  8. 2019Deep Eyedentification: Biometric Identification using Micro-Movements of the Eyecited by 25
  9. 2016Improving sentence compression by learning to predict gazecited by 24
  10. 2023Eyettention: An Attention-based Dual-Sequence Model for Predicting Human Scanpaths during Readingcited by 19
  11. 2021Towards Incremental Transformers: An Empirical Analysis of Transformer Models for Incremental NLUcited by 17
  12. 2020Incremental Processing in the Age of Non-Incremental Encoders: An Empirical Assessment of Bidirectional Models for Incremental NLUcited by 14
6 more
  1. 2021A Survey on Using Gaze Behaviour for Natural Language Processingcited by 11
  2. 2018Incremental Natural Language Processing: Challenges, Strategies, and Evaluationcited by 8
  3. 2018A Discriminative Model for Identifying Readers and Assessing Text Comprehension from Eye Movementscited by 8
  4. 2023TAPIR: Learning Adaptive Revision for Incremental Natural Language Understanding with a Two-Pass Modelcited by 5
  5. 2023Efficient Encoders for Streaming Sequence Taggingcited by 4
  6. 2023The Road to Quality is Paved with Good Revisions: A Detailed Evaluation Methodology for Revision Policies in Incremental Sequence Labellingcited by 2

Built on by

Not cited yet.