Imitation Learning for Neural Morphological String Transduction
This work addresses inefficiencies in morphological processing for NLP applications, though it is incremental in improving existing methods.
The paper tackles the problem of training neural transition-based string transducers for morphological tasks like inflection generation and lemmatization, achieving state-of-the-art performance on several benchmarks by eliminating the need for external character aligners or warm starts.
We employ imitation learning to train a neural transition-based string transducer for morphological tasks such as inflection generation and lemmatization. Previous approaches to training this type of model either rely on an external character aligner for the production of gold action sequences, which results in a suboptimal model due to the unwarranted dependence on a single gold action sequence despite spurious ambiguity, or require warm starting with an MLE model. Our approach only requires a simple expert policy, eliminating the need for a character aligner or warm start. It also addresses familiar MLE training biases and leads to strong and state-of-the-art performance on several benchmarks.