id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_ms7apx6u4vd7vjvtuih7h6vqx4	Tatsunori B. Hashimoto	Word Embeddings as Metric Recovery in Semantic Spaces	2016	14	.pdf	application/pdf	8281	1046	66	Word Embeddings as Metric Recovery in Semantic Spaces as metric recovery of a semantic space unifies existing word embedding algorithms, ties a simple, principled, direct metric recovery algorithm that performs on par with the state-ofthe-art word embedding and manifold learning co-occurrences through semantic similarity assessments, and demonstrate that the observed cooccurrence counts indeed possess statistical properties that are consistent with an underlying Euclidean To this end, we unify existing word embedding algorithms as statistically consistent metric recovery study, unifying existing algorithms as consistent metric recovery methods based on cooccurrence counts from simple Markov random two sections, we establish this connection by framing word embedding algorithms that operate on cooccurrences as metric recovery methods. distances between words using the negative log cooccurrence counts (Section 3), while manifold learning approximates semantic distances using neighborhood graphs built from local comparisons of the 3. Use a word embedding method on this corpus to generate d-dimensional vector representations of the data.	./cache/work_ms7apx6u4vd7vjvtuih7h6vqx4.pdf	./txt/work_ms7apx6u4vd7vjvtuih7h6vqx4.txt