Generalized Measures of Anticipation and Responsivity in Online Language Processing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Giulianelli, Mario, Opedal, Andreas, Cotterell, Ryan
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917801152741376
author Giulianelli, Mario
Opedal, Andreas
Cotterell, Ryan
author_facet Giulianelli, Mario
Opedal, Andreas
Cotterell, Ryan
contents We introduce a generalization of classic information-theoretic measures of predictive uncertainty in online language processing, based on the simulation of expected continuations of incremental linguistic contexts. Our framework provides a formal definition of anticipatory and responsive measures, and it equips experimenters with the tools to define new, more expressive measures beyond standard next-symbol entropy and surprisal. While extracting these standard quantities from language models is convenient, we demonstrate that using Monte Carlo simulation to estimate alternative responsive and anticipatory measures pays off empirically: New special cases of our generalized formula exhibit enhanced predictive power compared to surprisal for human cloze completion probability as well as ELAN, LAN, and N400 amplitudes, and greater complementarity with surprisal in predicting reading times.
format Preprint
id arxiv_https___arxiv_org_abs_2409_10728
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Generalized Measures of Anticipation and Responsivity in Online Language Processing
Giulianelli, Mario
Opedal, Andreas
Cotterell, Ryan
Computation and Language
Artificial Intelligence
Information Theory
We introduce a generalization of classic information-theoretic measures of predictive uncertainty in online language processing, based on the simulation of expected continuations of incremental linguistic contexts. Our framework provides a formal definition of anticipatory and responsive measures, and it equips experimenters with the tools to define new, more expressive measures beyond standard next-symbol entropy and surprisal. While extracting these standard quantities from language models is convenient, we demonstrate that using Monte Carlo simulation to estimate alternative responsive and anticipatory measures pays off empirically: New special cases of our generalized formula exhibit enhanced predictive power compared to surprisal for human cloze completion probability as well as ELAN, LAN, and N400 amplitudes, and greater complementarity with surprisal in predicting reading times.
title Generalized Measures of Anticipation and Responsivity in Online Language Processing
topic Computation and Language
Artificial Intelligence
Information Theory
url https://arxiv.org/abs/2409.10728