Age of Exposure: A Model of Word Learning

Authors: Mihai Dascalu, Danielle McNamara, Scott Crossley, Stefan Trausan-Matu

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental This study describes a proof of concept based on the on a large-scale learning corpus (i.e., TASA). The results indicate that Ao E indices yield strong associations with human ratings of age of acquisition, word frequency, entropy, and human lexical response latencies providing evidence of convergent validity. and Pearson product moment correlations (see Table 2) demonstrated that all Ao E variables had strong (i.e., strong effects, r > .500) and significant relations (i.e., p < .001) with the selected convergent validity indices related to lexical sophistication and knowledge.
Researcher Affiliation Academia Mihai Dascalu1, Danielle S. Mc McNamara2, Scott Crossley3 and Stefan Trausan-Matu1 1University Politehnica of Bucharest, 313 Splaiul Indepententei, Bucharest, Romania 2Arizona State University, PO Box 872111, Tempe, AZ 85287, USA 3Georgia State University, 25 Park Place, Ste 1500, Atlanta, GA 30303, USA
Pseudocode No The paper describes the steps of its method in prose and bullet points, but does not include formal pseudocode or algorithm blocks.
Open Source Code No The paper does not provide concrete access information (e.g., a specific repository link or an explicit statement of code release) for the methodology described.
Open Datasets Yes Our Ao E indices were implemented using the TASA corpus (http://lsa.colorado.edu/spaces.html) which was segmented based on Degrees of Reading Power (DRP; (Koslin et al., 1987) into 13 grade levels (Mc Namara, Graesser, & Louwerse, 2012).
Dataset Splits Yes Our Ao E indices were implemented using the TASA corpus (http://lsa.colorado.edu/spaces.html) which was segmented based on Degrees of Reading Power (DRP; (Koslin et al., 1987) into 13 grade levels (Mc Namara, Graesser, & Louwerse, 2012). The n-th Ao E intermediate model contained all the documents of complexity 1 up to n (with a corresponding notation of [1 n]). and Table 1. Statistics of intermediate and mature models after lemmatization and stop-words elimination.
Hardware Specification No The paper does not provide specific hardware details (e.g., CPU, GPU models, or memory specifications) used for running the experiments.
Software Dependencies No The paper mentions using 'Stanford Core NLP Morph Annotator' and 'Snowball list' for preprocessing, but does not provide specific version numbers for these or any other software dependencies.
Experiment Setup Yes For the most developed model, we opted to use 100 topics as indicated by Blei (2012). and experimentally, a threshold of .4 provided the best results when considering all possible thresholds from 0.4 to 0.7 with a 0.1 increment. and a 0.4 threshold provided the best results.