Approximate Structured Diffusion for Sequence Labelling
It addresses the limitation of finite decision span in CRFs for sequence labelling tasks, improving performance on long-range dependencies.
The paper introduces an approximate structured diffusion method for sequence labelling that conditions a CRF on noisy versions of the entire label sequence, achieving a 16.5% error reduction in POS-tagging.
Sequence labelling, a core task of Natural Language Processing (NLP), consists in assigning each token of an input sentence a label. From a Machine Learning point of view, sequence labelling is often cast as a Linear-Chain Conditional Random Field (CRF) parametrised by a neural network. While this approach gives good empirical results, CRFs assume a finite decision span (eg label bigrams) which can limit their expressivity and hurt performance when long-range dependencies are required. We show we can leverage diffusion to train a CRF conditioned on an entire label sequence, with the caveat that the condition is on a noisy version of labels. We show experimentally that this method, in conjunction with approximate CRF inference, improves label accuracy with a 16.5% error reduction for POS-tagging.