Supertagging with Factorial Hidden Markov Models

Factorial Hidden Markov Models (FHMM) support joint inference for multiple sequence prediction tasks. Here, we use them to jointly predict part-of-speech tag and supertag sequences with varying levels of supervision. We show that supervised training of FHMM models improves performance compared to standard HMMs, especially when labeled training data is scarce. Secondly, we show that an FHMM and a maximum entropy Markov model in a single step co-training setup improves the performance of both models when there is limited labeled training data. Finally, we find that FHMMs trained from tag dictionaries rather than labeled examples also perform better than a standard HMM.