id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_7drikjhpivhsjj7hgjbx3kyriy	Karl Stratos	Unsupervised Part-Of-Speech Tagging with Anchor Hidden Markov Models	2016	14	.pdf	application/pdf	7681	1082	74	Unsupervised Part-Of-Speech Tagging with Anchor Hidden Markov Models These HMMs impose an assumption that each hidden state is associated with an observation state ("anchor word") that can appear under no other state. Because each hidden state is associated with an observation, we can examine the set of such anchor observations to qualitatively evaluate the learned model. In Section 3, we define the model family of anchor HMMs. In Section 4, we derive a matrix decomposition algorithm for estimating the parameters of an anchor A concrete algorithm for factorizing a matrix satisfying Condition 4.1 is given in Figure 1 (Arora model, the algorithm Learn-Anchor-HMM in Figure 2 outputs (π,T,O) up to a permutation on hidden states. We speculate that this is because a Brown model is rather appropriate for the POS tagging task; many words are The anchor condition corresponds to assuming that each POS tag has at least one word that	./cache/work_7drikjhpivhsjj7hgjbx3kyriy.pdf	./txt/work_7drikjhpivhsjj7hgjbx3kyriy.txt