Hidden Markov Chains, Entropic Forward-Backward, and Part-Of-Speech Tagging
This addresses a long-standing problem in NLP for researchers and practitioners by reviving HMC as a competitive method for sequential tasks, though it appears incremental as it modifies existing restoration algorithms rather than introducing a new paradigm.
The paper tackles the limitation of Hidden Markov Chains (HMC) in handling arbitrary features for sequential data like in NLP, by introducing Entropic Forward-Backward (EFB) probabilities, which enable HMC to incorporate features similarly to Maximum Entropy Markov Models (MEMM) and show superiority in Part-Of-Speech Tagging.
The ability to take into account the characteristics - also called features - of observations is essential in Natural Language Processing (NLP) problems. Hidden Markov Chain (HMC) model associated with classic Forward-Backward probabilities cannot handle arbitrary features like prefixes or suffixes of any size, except with an independence condition. For twenty years, this default has encouraged the development of other sequential models, starting with the Maximum Entropy Markov Model (MEMM), which elegantly integrates arbitrary features. More generally, it led to neglect HMC for NLP. In this paper, we show that the problem is not due to HMC itself, but to the way its restoration algorithms are computed. We present a new way of computing HMC based restorations using original Entropic Forward and Entropic Backward (EFB) probabilities. Our method allows taking into account features in the HMC framework in the same way as in the MEMM framework. We illustrate the efficiency of HMC using EFB in Part-Of-Speech Tagging, showing its superiority over MEMM based restoration. We also specify, as a perspective, how HMCs with EFB might appear as an alternative to Recurrent Neural Networks to treat sequential data with a deep architecture.