A Latent-Variable Model for Intrinsic Probing
This work addresses the need for more precise analysis of linguistic information in NLP representations, though it is incremental as it builds on existing intrinsic probing techniques.
The paper tackles the problem of identifying where linguistic attributes are encoded in pre-trained representations by proposing a latent-variable model for intrinsic probing, resulting in tighter mutual information estimates than previous methods and empirical evidence of cross-lingually entangled morphosyntax.
The success of pre-trained contextualized representations has prompted researchers to analyze them for the presence of linguistic information. Indeed, it is natural to assume that these pre-trained representations do encode some level of linguistic knowledge as they have brought about large empirical improvements on a wide variety of NLP tasks, which suggests they are learning true linguistic generalization. In this work, we focus on intrinsic probing, an analysis technique where the goal is not only to identify whether a representation encodes a linguistic attribute but also to pinpoint where this attribute is encoded. We propose a novel latent-variable formulation for constructing intrinsic probes and derive a tractable variational approximation to the log-likelihood. Our results show that our model is versatile and yields tighter mutual information estimates than two intrinsic probes previously proposed in the literature. Finally, we find empirical evidence that pre-trained representations develop a cross-lingually entangled notion of morphosyntax.