A Semantics-based Model for Predicting Children's Vocabulary

Authors: Ishaan Grover, Hae Won Park, Cynthia Breazeal

IJCAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We show that the proposed semantics-based model outperforms models that do not use word semantics (semantics-free models) on average. A subject-level analysis of results reveals that different models perform well for different children, thus motivating the need to combine predictions. To this end, we use two methods to combine predictions from semantics-based and semantics-free models and show that these methods yield better predictions of a child s vocabulary knowledge. Our results motivate the use of semantics-based models to assess children s vocabulary knowledge and build ITS that maximizes children s semantic understanding of words.
Researcher Affiliation Academia Ishaan Grover , Hae Won Park and Cynthia Breazeal Massachusetts Institute of Technology 77 Massachusetts Avenue, Cambridge, Massachusetts, 02139 {igrover, haewon, breazeal}@mit.edu
Pseudocode No No pseudocode or algorithm blocks are provided in the paper.
Open Source Code No The paper does not provide any concrete access to source code for the methodology described.
Open Datasets Yes We used the dataset from our work in [Park et al., 2019].
Dataset Splits No The paper describes the dataset used (from [Park et al., 2019]) and how some data was excluded, but does not provide explicit train/validation/test dataset splits (e.g., percentages, counts, or references to standard predefined splits).
Hardware Specification No No specific hardware details (e.g., GPU/CPU models, memory specifications) used for running experiments are mentioned in the paper.
Software Dependencies No The paper mentions techniques and external resources (e.g., GloVe word vectors, SUBTLEXus database) but does not provide specific ancillary software names with version numbers (e.g., Python 3.8, PyTorch 1.9) used for replication.
Experiment Setup Yes We define semantic distance between two words as the cosine distance between their pre-trained common crawl Glo Ve word vectors (300 dimensional) [Pennington et al., 2014]. Two words with vector representations v1 and v2 are said to be semantically similar if cos(v1, v2) ϵ (after manually testing different values of semantic similarity on different words, we set ϵ = 0.6 for this study). The zipf scale converts word frequencies (per billion words) into a log-based scale with values between 1-7 and is independent of the size of the corpus used [Van Heuven et al., 2014]. Then for each child, we train a personalized logistic regression model using zipf score of words in Wpre as training data and whether or not a child got the words correct as training labels (binary classification). Since there is no prior informing which of the two distributions should be given a higher weight, we assume an equal weight (0.5) for each of the distributions and compute the posterior by taking the weighted sum of distributions from the two models.