Binary Encoded Word Mover’s Distance Article Swipe
Related Concepts
Word (group theory)
Metric (unit)
Computer science
Similarity (geometry)
Binary number
Earth mover's distance
Semantic similarity
Computation
Point (geometry)
Artificial intelligence
Natural language processing
Algorithm
Theoretical computer science
Arithmetic
Mathematics
Geometry
Operations management
Economics
Image (mathematics)
C. A. Johnson
·
YOU?
·
· 2022
· Open Access
·
· DOI: https://doi.org/10.18653/v1/2022.repl4nlp-1.17
· OA: W4285226496
YOU?
·
· 2022
· Open Access
·
· DOI: https://doi.org/10.18653/v1/2022.repl4nlp-1.17
· OA: W4285226496
Word Mover’s Distance is a textual distance metric which calculates the minimum transport cost between two sets of word embeddings. This metric achieves impressive results on semantic similarity tasks, but is slow and difficult to scale due to the large number of floating point calculations. This paper demonstrates that by combining pre-existing lower bounds with binary encoded word vectors, the metric can be rendered highly efficient in terms of computation time and memory while still maintaining accuracy on several textual similarity tasks.
Related Topics
Finding more related topics…