Adaptive Forgetting Curves for Spaced Repetition Language Learning
Project Overview
The document examines the integration of generative AI in education, particularly through the development of adaptive forgetting curves for spaced repetition in language learning. It highlights the use of intelligent tutoring systems that leverage various forgetting curve models, incorporating psychological and linguistic features to enhance vocabulary recall for English as a second language learners. Key findings indicate that word complexity significantly affects memory retention, suggesting that tailored learning experiences can optimize vocabulary acquisition. The implementation of neural networks is shown to improve the modeling of these complex relationships, leading to more effective educational tools. Overall, the research underscores the potential of generative AI to transform language learning by providing personalized and adaptive learning pathways that cater to individual needs, ultimately enhancing educational outcomes.
Key Applications
Adaptive Forgetting Curves for Spaced Repetition Language Learning
Context: Language learning for English as a second language learners through an online vocabulary teaching platform.
Implementation: Utilized the Duolingo spaced repetition dataset to train and evaluate various forgetting curve models, including modifications to the half-life regression model and the introduction of neural networks.
Outcomes: Improved modeling of vocabulary recall with significant improvements in prediction accuracy through the incorporation of word complexity as a feature.
Challenges: Single-dimensional capture of user behavior and limitations of certain features like SUBTLEX in adequately representing learner experiences.
Implementation Barriers
Technical Barrier
Challenges in accurately capturing user behavior and the complexity of language learning.
Proposed Solutions: Future work includes incorporating high-dimensional user embeddings to better capture user-specific signals.
Project Team
Ahmed Zaidi
Researcher
Andrew Caines
Researcher
Russell Moore
Researcher
Paula Buttery
Researcher
Andrew Rice
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Ahmed Zaidi, Andrew Caines, Russell Moore, Paula Buttery, Andrew Rice
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai