A cross-situational word learning framework
Scilab Python
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
core
data
starter
.gitignore
README.md
__init__.py

README.md

word_learning

This code provides a framework for modeling cross-situational word learning. The core algorithm implements the model of Fazly et al. (2010), which is an incremental and probabilistic word learner.

The code also includes extensions of the above model that allow investigation of:

  • Individual differences in word learning. (Nematzadeh et al., 2011, 2012a & 2014a)
  • The role of memory and attention in word learning. (Nematzadeh et al., 2012b & 2013)
  • The acquisition of a semantic network. (Nematzadeh et al., 2014b)

An extension of this model has been used to study novel word generalization (Nematzadeh et al., 2015); the code can be found here.

References:

Starter code is provided in starter/main.py, and development and test data are located at data/input_wn_fu_cs_scaled_categ.dev and data/input_wn_fu_cs_scaled_categ.tst, respectively. The gold standard lexicon, which was used to generate the dev/test data, and which can be used to compute metrics such as the acquisition score, is located at data/all_catf_norm_prob_lexicon_cs.all.

Requirements: Python 2, numpy, scipy