skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


Title: Incorporating Word-level Phonemic Decoding into Readability Assessment
Current approaches in automatic readability assessment have found success with the use of large language models and transformer architectures. These techniques lead to accuracy improvement, but they do not offer the interpretability that is uniquely required by the audience most often employing readability assessment tools: teachers and educators. Recent work that employs more traditional machine learning methods has highlighted the linguistic importance of considering semantic and syntactic characteristics of text in readability assessment by utilizing handcrafted feature sets. Research in Education suggests that, in addition to semantics and syntax, phonetic and orthographic instruction are necessary for children to progress through the stages of reading and spelling development; children must first learn to decode the letters and symbols on a page to recognize words and phonemes and their connection to speech sounds. Here, we incorporate this word-level phonemic decoding process into readability assessment by crafting a phonetically-based feature set for grade-level classification for English. Our resulting feature set shows comparable performance to much larger, semantically- and syntactically-based feature sets, supporting the linguistic value of orthographic and phonetic considerations in readability assessment.  more » « less
Award ID(s):
1763649
PAR ID:
10513286
Author(s) / Creator(s):
; ; ; ;
Publisher / Repository:
Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)
Date Published:
Page Range / eLocation ID:
8998–9009
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ; (, ISCA INTERSPEECH-2023)
    N/A (Ed.)
    Automatic pronunciation assessment (APA) plays an important role in providing feedback for self-directed language learners in computer-assisted pronunciation training (CAPT). Several mispronunciation detection and diagnosis (MDD) systems have achieved promising performance based on end-to-end phoneme recognition. However, assessing the intelligibility of second language (L2) remains a challenging problem. One issue is the lack of large-scale labeled speech data from non-native speakers. Additionally, relying only on one aspect (e.g., accuracy) at a phonetic level may not provide a sufficient assessment of pronunciation quality and L2 intelligibility. It is possible to leverage segmental/phonetic-level features such as goodness of pronunciation (GOP), however, feature granularity may cause a discrepancy in prosodic-level (suprasegmental) pronunciation assessment. In this study, Wav2vec 2.0-based MDD and Goodness Of Pronunciation feature-based Transformer are employed to characterize L2 intelligibility. Here, an L2 speech dataset, with human-annotated prosodic (suprasegmental) labels, is used for multi-granular and multi-aspect pronunciation assessment and identification of factors important for intelligibility in L2 English speech. The study provides a transformative comparative assessment of automated pronunciation scores versus the relationship between suprasegmental features and listener perceptions, which taken collectively can help support the development of instantaneous assessment tools and solutions for L2 learners. 
    more » « less
  2. ; ; ; (, Autism research)
    Some autistic children acquire foreign languages from exposure to screens. Such Unexpected Bilingualism (UB) is therefore not driven by social interaction; rather, language acquisition appears to rely on less socially mediated learning and other cognitive processes. We hypothesize that UB children may rely on other cues, such as acoustic cues, of the linguistic input. Previous research indicates enhanced pitch processing in some autistic children, often associated with language delays and difficulties in forming stable phonological categories due to sensitivity to subtle linguistic variations. We propose that repetitive screen-based input simplifies linguistic complexity, allowing focus on individual cues. This study hypothesizes that autistic UB children exhibit superior pitch discrimination compared to both autistic and non-autistic peers. From a sample of 46 autistic French-speaking children aged 9 to 16, 12 were considered as UB. These children, along with 45 non-autistic children, participated in a two-alternative forced-choice pitch discrimination task. They listened to pairs of pure tones, 50% of which differed by 3% (easy), 2% (medium), or 1% (hard). A stringent comparison of performance revealed that only the autistic UB group performed above chance for tone pairs that differed, across all conditions. This group demonstrated superior pitch discrimination relative to autistic and non-autistic peers. This study establishes the phenomenon of UB in autism and provides evidence for enhanced pitch discrimination in this group. Acute perception of auditory information, combined with repeated language content, may facilitate UB children's focus on phonetic features, and help acquire a language with no communicative support or motivation. 
    more » « less
  3. ; (, Language documentation and conservation)
    We present an initial web-based tool for St. Lawrence Island/Central Siberian Yupik, an endangered language of Alaska and Russia. This work is supported by the local language community on St. Lawrence Island, and includes an orthographic utility to convert from standard Latin orthography into a fully transparent representation, a preliminary spell checker, a Latin-to-Cyrillic transliteration tool, and a preliminary Cyrillic-to-Latin transliteration tool. Also included is a utility to convert from standard Latin orthography into both IPA and Americanist phonetic notation. Our utility is also capable of explicitly marking syllable boundaries and stress in the standard Latin orthography using the conventions of Jacobson (2001), as well as in Cyrillic and in standard IPA notation. These tools are designed to facilitate the digitization of existing Yupik resources, facilitate additional linguistic field work, and most importantly, bolster efforts by the local Yupik communities in the U.S. and in Russia to promote Yupik usage and literacy, especially among Yupik youth. 
    more » « less
  4. ; ; (, Educational Assessment)
    Little is known about mismatches between the language of mathematics testing instruments and the rich linguistic repertoires that African American children develop at home and in the community. The current study aims to provide a proof of concept and novel explanatory item response design that uses error analysis to investigate the relationship between AAE child language and chil- dren’s mathematics assessment outcomes. Here, we illustrate 2nd and 3rd grade children’s qualitative patterns of performance on arithmetic tasks in relation to their AAE dialect use and elaborate a unified framework for examining child and item level linguistic characteristics. Results suggest that children draw upon their emerging (bi)dialectal repertoire with arithmetic problems when selecting appropriate problem-solving strategies on language-formatted problems. The mismatch of assessment language formatting with children’s repertoires may disadvantage AAE speakers’ strategy selections and result in a language-based performance disadvantage unrelated to mathematical ability. 
    more » « less
  5. ; ; (, Proceedings of the Annual Conference of the Cognitive Science Society)
    Listeners draw on their knowledge of phonetic categories when identifying speech sounds, extracting meaningful structural features from auditory cues. We use a Bayesian model to investigate the extent to which their perceptions of linguistic content incorporate their full knowledge of the phonetic category structure, or only certain aspects of this knowledge. Simulations show that listeners are best modeled as attending primarily to the most salient phonetic feature of a category when interpreting a cue, possibly attending to other features only in cases of high ambiguity. These results support the conclusion that listeners ignore potentially informative correlations in favor of efficient communication. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email