Language Acquisition Research Tools

Code & data for the EMNLP 2024 paper: Is Child-Directed Speech Effective Training Data for Language Models?

A list of papers that you should probably read before you try to create artificial agents that can learn a language or a concept.

Depth-bounded grammar Induction Model with Inside-sampling

Lessons for learning a foreign language. Formerly a wiki, now a static site built with Hugo.

A curated repository of resources for Linguistics, the scientific study of language

This project was originally my term project for a computational linguistics course at Pitt. It was turned into a research project later and I am working on publishing the work.

Guidelines, documentation, instructions, issues, etc.

Code for the paper "Sampling strategies in Siamese Networks for unsupervised speech representation learning" Interspeech 2018

"Flashcardify-GPT is a Python script that uses the OpenAI GPT-3.5 language model to generate language flashcards for learners. With Flashcardify-GPT, users can generate customized decks of flashcards with common phrases and vocabulary.

Code for the paper "Deep neural networks and humans both benefit from compositional structure"

Embodied agent that learns from human teaching; hybrid agent architecture with computer vision, AI planning, knowledge graph reasoning & generalization , inverse kinematics. Supported by the DARPA GAILA program.

Mora (formerly YouLingo) is an app that makes authentic language learning happen through user-generated short-form videos. This project was originally created for the UofTHacks 12 Hackathon and was the winner of Best University of Toronto Hack. It is currently in development with the team at Roi Digital.

Research code for exploring co-occurrence structure of nouns in child-directed speech

Code and material for the paper "Beyond the Self: Using Grounded Affordances to Interpret and Describe Others' Actions" (doi: 10.1109/TCDS.2018.2882140)

A replication of an experiment by Reali and Christiansen (2005) disputing the basic assumptions of Chomsky's Poverty of Stimulus theory.

Pipeline for the ACQDIV Corpus Database

a list of links about Automatic Language Growth