Abstract:We present StarSpace, a general-purpose neural embedding model that can solve a wide variety of problems: labeling tasks such as text classification, ranking tasks such as information retrieval/web search, collaborative filtering-based or content-based recommendation, embedding of multi-relational graphs, and learning word, sentence or document level embeddings. In each case the model works by embedding those entities comprised of discrete features and comparing them against each other -- learning similarities dependent on the task. Empirical results on a number of tasks show that StarSpace is highly competitive with existing methods, whilst also being generally applicable to new cases where those methods are not.
Submission history
From: Jason Weston [view email]
[v1]
Tue, 12 Sep 2017 14:16:56 UTC (26 KB)
[v2]
Wed, 13 Sep 2017 12:19:23 UTC (26 KB)
[v3]
Thu, 14 Sep 2017 13:06:43 UTC (26 KB)
[v4]
Tue, 26 Sep 2017 15:00:56 UTC (26 KB)
[v5]
Tue, 21 Nov 2017 02:59:57 UTC (28 KB)