Meta-Learning a Dynamical Language Model

Thomas Wolf, Julien Chaumond, Clement Delangue

Abstract: We consider the task of word-level language modeling and study the possibility of combining hidden-states-based short-term representations with medium-term representations encoded in dynamical weights of a language model. Our work extends recent experiments on language models with dynamically evolving weights by casting the language modeling problem into an online learning-to-learn framework in which a meta-learner is trained by gradient-descent to continuously update a language model weights.

Keywords: language models, hierarchical representations, meta-learning, nonstationarity, catastrophic forgetting, recurrent neural network

TL;DR: Language modeling with dynamical weights as an instance of continuous meta-learning