Interpretable Mnemonic Generation for Kanji Learning via Expectation-Maximization
This addresses the challenge of script differences for learners of Japanese from Roman alphabet backgrounds by providing interpretable mnemonics, though it is incremental as it builds on existing mnemonic strategies with a novel algorithmic approach.
The paper tackles the problem of generating interpretable keyword mnemonics for learning kanji by proposing a generative framework that models mnemonic construction with common rules learned via an Expectation-Maximization algorithm, resulting in effective performance in cold-start settings for new learners.
Learning Japanese vocabulary is a challenge for learners from Roman alphabet backgrounds due to script differences. Japanese combines syllabaries like hiragana with kanji, which are logographic characters of Chinese origin. Kanji are also complicated due to their complexity and volume. Keyword mnemonics are a common strategy to aid memorization, often using the compositional structure of kanji to form vivid associations. Despite recent efforts to use large language models (LLMs) to assist learners, existing methods for LLM-based keyword mnemonic generation function as a black box, offering limited interpretability. We propose a generative framework that explicitly models the mnemonic construction process as driven by a set of common rules, and learn them using a novel Expectation-Maximization-type algorithm. Trained on learner-authored mnemonics from an online platform, our method learns latent structures and compositional rules, enabling interpretable and systematic mnemonics generation. Experiments show that our method performs well in the cold-start setting for new learners while providing insight into the mechanisms behind effective mnemonic creation.