Speeding up Word Mover's Distance and its variants via properties of distances between embeddings
Matheus Werner, Eduardo Laber
Code Available — Be the first to reproduce this paper.
ReproduceCode
- github.com/matwerner/fast-wmdOfficialIn papernone★ 0
Abstract
The Word Mover's Distance (WMD) proposed by Kusner et al. is a distance between documents that takes advantage of semantic relations among words that are captured by their embeddings. This distance proved to be quite effective, obtaining state-of-art error rates for classification tasks, but is also impracticable for large collections/documents due to its computational complexity. For circumventing this problem, variants of WMD have been proposed. Among them, Relaxed Word Mover's Distance (RWMD) is one of the most successful due to its simplicity, effectiveness, and also because of its fast implementations. Relying on assumptions that are supported by empirical properties of the distances between embeddings, we propose an approach to speed up both WMD and RWMD. Experiments over 10 datasets suggest that our approach leads to a significant speed-up in document classification tasks while maintaining the same error rates.
Tasks
Benchmark Results
| Dataset | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| Amazon | REL-RWMD k-NN | Accuracy | 93.03 | — | Unverified |
| BBCSport | REL-RWMD k-NN | Accuracy | 95.18 | — | Unverified |
| Classic | REL-RWMD k-NN | Accuracy | 96.85 | — | Unverified |
| Recipe | REL-RWMD k-NN | Accuracy | 56.8 | — | Unverified |
| Reuters-21578 | REL-RWMD k-NN | Accuracy | 95.61 | — | Unverified |
| REL-RWMD k-NN | Accuracy | 71.05 | — | Unverified |