NU HLT at CMCL 2022 Shared Task: Multilingual and Crosslingual Prediction of Human Reading Behavior in Universal Language Space

Joseph Marvin Imperial

doi:10.18653/v1/2022.cmcl-1.12

NU HLT at CMCL 2022 Shared Task: Multilingual and Crosslingual Prediction of Human Reading Behavior in Universal Language Space

National University, Philippines

Research output: Chapter or section in a book/report/conference proceeding › Chapter in a published conference proceeding

2 Citations (SciVal)

Abstract

In this paper, we present a unified model that works for both multilingual and crosslingual prediction of reading times of words in various languages. The secret behind the success of this model is in the preprocessing step where all words are transformed to their universal language representation via the International Phonetic Alphabet (IPA). To the best of our knowledge, this is the first study to favorably exploit this phonological property of language for the two tasks. Various feature types were extracted covering basic frequencies, n-grams, information theoretic, and psycholinguistically-motivated predictors for model training. A finetuned Random Forest model obtained best performance for both tasks with 3.8031 and 3.9065 MAE scores for mean first fixation duration (FFDAvg) and mean total reading time (TRTAvg) respectively.

Original language	English
Title of host publication	CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop
Editors	Emmanuele Chersoni, Nora Hollenstein, Cassandra L. Jacobs, Yohei Oseki, Laurent Prevot, Enrico Santus
Publisher	Association for Computational Linguistics (ACL)
Pages	108-113
Number of pages	6
ISBN (Electronic)	9781955917292
DOIs	https://doi.org/10.18653/v1/2022.cmcl-1.12
Publication status	Published - 31 Dec 2022
Event	12th Workshop on Cognitive Modeling and Computational Linguistics, CMCL 2022 - Dublin, Ireland Duration: 26 May 2022 → …

Publication series

Name	CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop

Conference

Conference	12th Workshop on Cognitive Modeling and Computational Linguistics, CMCL 2022
Country/Territory	Ireland
City	Dublin
Period	26/05/22 → …

Bibliographical note

Publisher Copyright:
© 2022 Association for Computational Linguistics.

ASJC Scopus subject areas

Language and Linguistics
Artificial Intelligence
Software
Linguistics and Language

Access to Document

10.18653/v1/2022.cmcl-1.12Licence: CC BY

Cite this

Imperial, J. M. (2022). NU HLT at CMCL 2022 Shared Task: Multilingual and Crosslingual Prediction of Human Reading Behavior in Universal Language Space. In E. Chersoni, N. Hollenstein, C. L. Jacobs, Y. Oseki, L. Prevot, & E. Santus (Eds.), CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop (pp. 108-113). (CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop). Association for Computational Linguistics (ACL). https://doi.org/10.18653/v1/2022.cmcl-1.12

NU HLT at CMCL 2022 Shared Task: Multilingual and Crosslingual Prediction of Human Reading Behavior in Universal Language Space. / Imperial, Joseph Marvin.
CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop. ed. / Emmanuele Chersoni; Nora Hollenstein; Cassandra L. Jacobs; Yohei Oseki; Laurent Prevot; Enrico Santus. Association for Computational Linguistics (ACL), 2022. p. 108-113 (CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop).

Research output: Chapter or section in a book/report/conference proceeding › Chapter in a published conference proceeding

Imperial, JM 2022, NU HLT at CMCL 2022 Shared Task: Multilingual and Crosslingual Prediction of Human Reading Behavior in Universal Language Space. in E Chersoni, N Hollenstein, CL Jacobs, Y Oseki, L Prevot & E Santus (eds), CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop. CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop, Association for Computational Linguistics (ACL), pp. 108-113, 12th Workshop on Cognitive Modeling and Computational Linguistics, CMCL 2022, Dublin, Ireland, 26/05/22. https://doi.org/10.18653/v1/2022.cmcl-1.12

Imperial JM. NU HLT at CMCL 2022 Shared Task: Multilingual and Crosslingual Prediction of Human Reading Behavior in Universal Language Space. In Chersoni E, Hollenstein N, Jacobs CL, Oseki Y, Prevot L, Santus E, editors, CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop. Association for Computational Linguistics (ACL). 2022. p. 108-113. (CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop). doi: 10.18653/v1/2022.cmcl-1.12

Imperial, Joseph Marvin. / NU HLT at CMCL 2022 Shared Task : Multilingual and Crosslingual Prediction of Human Reading Behavior in Universal Language Space. CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop. editor / Emmanuele Chersoni ; Nora Hollenstein ; Cassandra L. Jacobs ; Yohei Oseki ; Laurent Prevot ; Enrico Santus. Association for Computational Linguistics (ACL), 2022. pp. 108-113 (CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop).

@inproceedings{fe62a22d6c4d4dd7b3ff0060664fc59f,

title = "NU HLT at CMCL 2022 Shared Task: Multilingual and Crosslingual Prediction of Human Reading Behavior in Universal Language Space",

abstract = "In this paper, we present a unified model that works for both multilingual and crosslingual prediction of reading times of words in various languages. The secret behind the success of this model is in the preprocessing step where all words are transformed to their universal language representation via the International Phonetic Alphabet (IPA). To the best of our knowledge, this is the first study to favorably exploit this phonological property of language for the two tasks. Various feature types were extracted covering basic frequencies, n-grams, information theoretic, and psycholinguistically-motivated predictors for model training. A finetuned Random Forest model obtained best performance for both tasks with 3.8031 and 3.9065 MAE scores for mean first fixation duration (FFDAvg) and mean total reading time (TRTAvg) respectively.",

author = "Imperial, \{Joseph Marvin\}",

note = "Publisher Copyright: {\textcopyright} 2022 Association for Computational Linguistics.; 12th Workshop on Cognitive Modeling and Computational Linguistics, CMCL 2022 ; Conference date: 26-05-2022",

year = "2022",

month = dec,

day = "31",

doi = "10.18653/v1/2022.cmcl-1.12",

language = "English",

series = "CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop",

publisher = "Association for Computational Linguistics (ACL)",

pages = "108--113",

editor = "Emmanuele Chersoni and Nora Hollenstein and Jacobs, \{Cassandra L.\} and Yohei Oseki and Laurent Prevot and Enrico Santus",

booktitle = "CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop",

}

TY - GEN

T1 - NU HLT at CMCL 2022 Shared Task

T2 - 12th Workshop on Cognitive Modeling and Computational Linguistics, CMCL 2022

AU - Imperial, Joseph Marvin

PY - 2022/12/31

Y1 - 2022/12/31

N2 - In this paper, we present a unified model that works for both multilingual and crosslingual prediction of reading times of words in various languages. The secret behind the success of this model is in the preprocessing step where all words are transformed to their universal language representation via the International Phonetic Alphabet (IPA). To the best of our knowledge, this is the first study to favorably exploit this phonological property of language for the two tasks. Various feature types were extracted covering basic frequencies, n-grams, information theoretic, and psycholinguistically-motivated predictors for model training. A finetuned Random Forest model obtained best performance for both tasks with 3.8031 and 3.9065 MAE scores for mean first fixation duration (FFDAvg) and mean total reading time (TRTAvg) respectively.

AB - In this paper, we present a unified model that works for both multilingual and crosslingual prediction of reading times of words in various languages. The secret behind the success of this model is in the preprocessing step where all words are transformed to their universal language representation via the International Phonetic Alphabet (IPA). To the best of our knowledge, this is the first study to favorably exploit this phonological property of language for the two tasks. Various feature types were extracted covering basic frequencies, n-grams, information theoretic, and psycholinguistically-motivated predictors for model training. A finetuned Random Forest model obtained best performance for both tasks with 3.8031 and 3.9065 MAE scores for mean first fixation duration (FFDAvg) and mean total reading time (TRTAvg) respectively.

UR - https://www.scopus.com/pages/publications/85137075754

U2 - 10.18653/v1/2022.cmcl-1.12

DO - 10.18653/v1/2022.cmcl-1.12

M3 - Chapter in a published conference proceeding

AN - SCOPUS:85137075754

T3 - CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop

SP - 108

EP - 113

BT - CMCL 2022 - Workshop on Cognitive Modeling and Computational Linguistics, Proceedings of the Workshop

A2 - Chersoni, Emmanuele

A2 - Hollenstein, Nora

A2 - Jacobs, Cassandra L.

A2 - Oseki, Yohei

A2 - Prevot, Laurent

A2 - Santus, Enrico

PB - Association for Computational Linguistics (ACL)

Y2 - 26 May 2022

ER -

NU HLT at CMCL 2022 Shared Task: Multilingual and Crosslingual Prediction of Human Reading Behavior in Universal Language Space

Abstract

Publication series

Conference

Bibliographical note

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this