One Size Does Not Fit All: The Case for Personalised Word Complexity Models

Abstract

Complex Word Identification (CWI) aims to detect words within a text that a reader may find difficult to understand. It has been shown that CWI systems can improve text simplification, readability prediction and vocabulary acquisition modelling. However, the difficulty of a word is a highly idiosyncratic notion that depends on a reader's first language, proficiency and reading experience. In this paper, we show that personal models are best when predicting word complexity for individual readers. We use a novel active learning framework that allows models to be tailored to individuals and release a dataset of complexity annotations and models as a benchmark for further research.

A Hybrid Personalized Text Simplification Framework Leveraging the Deep Learning-based Transformer Model for Dyslexic Students

“…In the meantime, research on lexical complexity classification at the word level by Gooding and Tragut [10] and Balyan et al, [11] has transformed the binary classification into a more thorough multi-class classification. The study by Gooding and Tragut [10] has classified the words into seven classes based on the English Common European Framework of Reference for Languages (CEFR), while the study by Balyan et al, [11] has classified the words into four classes: easy, medium, difficult, and very difficult.…”

“…In the meantime, research on lexical complexity classification at the word level by Gooding and Tragut [10] and Balyan et al, [11] has transformed the binary classification into a more thorough multi-class classification. The study by Gooding and Tragut [10] has classified the words into seven classes based on the English Common European Framework of Reference for Languages (CEFR), while the study by Balyan et al, [11] has classified the words into four classes: easy, medium, difficult, and very difficult. In terms of methodologies, Gooding and Tragut [10] implemented the unsupervised active learning agglomerative clustering method, which performed the clustering from the bottom up, whereas Balyan et al, [11] strongly utilized the Machine Learning approach coupled with the NLP.…”

“…The study by Gooding and Tragut [10] has classified the words into seven classes based on the English Common European Framework of Reference for Languages (CEFR), while the study by Balyan et al, [11] has classified the words into four classes: easy, medium, difficult, and very difficult. In terms of methodologies, Gooding and Tragut [10] implemented the unsupervised active learning agglomerative clustering method, which performed the clustering from the bottom up, whereas Balyan et al, [11] strongly utilized the Machine Learning approach coupled with the NLP. Both studies extracted similar word-level features: uncommon words, word familiarities, word frequencies, word imageabilities, the average number of syllables per word, and the average number of characters per word.…”

“…Apart from the extrinsic-intrinsic personalization, classification of text complexity at the beginning of the text simplification workflow has been implemented by several studies, also with the aim to support personalization [6][7][8][9][10][11]. Each sentence is labelled either as simple-to-read or difficultto-read based on scores measured by sentence comparisons.…”

Generating and authoring high-variability exercises from authentic texts

“…Pilán et al ( 2017) identify a number of criteria for good seed sentences, including well-formedness, context independence, linguistic complexity and additional structural and lexical criteria. While we address most of the structural criteria, such as negated or interrogative contexts, with the parameters exposed to users, we deliberately do not restrict seed sentence selection based on lexical criteria, which are often user-dependent and better targeted by a macro-adaptive algorithm in the target ILTS (Gooding and Tragut, 2022). Compliance with context independence will be more likely when the co-text option is activated and can be addressed manually in the subsequent workflow step.…”