StarSpace: Embed All The Things!
This provides a versatile tool for practitioners needing embedding solutions across various domains, though it is incremental as it builds on existing embedding paradigms.
The authors introduced StarSpace, a general-purpose neural embedding model that can solve diverse problems including text classification, information retrieval, recommendation, and graph embedding by learning task-dependent similarities between discrete entities. Empirical results showed it is highly competitive with existing methods across multiple tasks while being more generally applicable to new cases.
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.