This repository contains the code for the supplementary material of the article "Opening the black box of language acquisition".
It contains a number of python files used to perform the simulations in that paper.
- Raw_input.py contains the code for generating probabilistic context free grammars in the format used by the learning agent
- MyLearners.py contains the classes defining the different types of learners
Learning curves are produces using:
- LearningCurves.py (4 learning curves for a NVN language)
- LearningCurvesMD.py (learning curve for a language containing mono- and ditransitive verbs)
- LearningCurvesRel.py (learning curve for a language containing relative clauses)
- LearningCurvesComplNP.py (learning curve for a language with complex NPs)
Grammar snapshots are produced using:
- GrammarSnapshots.py producing 4 Excel files: QLearnerC2_revisedFinal.xlsx, QLearnerN2_revisedFinal.xlsx, RWQLearnerC2_revisedFinal.xlsx, and RWQLearnerN2_revisedFinal.xlsx
- GrammarSnapshotsMD.py producing RWQLearnerC_MD_revisedFinal.xlsx
- grammarSnapshots-rel.py producing RWQLearnerC_rel_revisedFinal.xlsx
- grammarSnapshotsComplNP.py producing RWQLearnerC_MDComplNP.xlsx
Sentences extracted for the langauge with relative clauses are stored in the sentencesRelClauses.xlsx and sentencesRelClauses1.xlsx
Sentences extracted for the Complex NP language are stored in the sentencesCompNP_paper.xlsx
Other files are not relevant for the paper and were generated in earlier phases of development.