| Abstract |
Semantic Vector Spaces(SVS’s) that are constructed using semantic vector embedding techniques have proven use- ful in learning semantic vector representations of data across multiple domains. An important application area enabled by such techniques is the capability to represent software ser- vices and the service workflows in which they are embedded as semantic vectors. We have shown in previous work that these vector representations have significant advantages over alternative schemes for decentralized workflow construction. State-of-the-art embedding approaches assume all data required to construct the semantic vector space is available centrally. However, in a coalition setting different coalition partners may not be willing to share the training data necessary to construct a common coalition semantic vector space. Hence semantic vectors representing similar services or workflows, but constructed from different training data by different coalition partners cannot be compared. In this paper we specifically focus on the issue of mapping semantic vectors between such coalition SVS’s, so that complementary services and workflows owned by other coalition partners required to achieve specific mission goals goals can be identified and used. We describe a number of different approaches for mapping between SVS’s and demonstrate how it is possible to do this for semantic vectors with real valued coefficients and for the high dimensional binary semantic vectors. |
| Authors |
- Graham Bent (IBM UK)
- Declan Millar (IBM UK)
- Douglas Summers-Stay (ARL)
- Shalisa Witherspoon (IBM US)
- Jae-Wook Ahn (IBM US)
|
| Date |
Sep-2020 |
| Venue |
4th Annual Fall Meeting of the DAIS ITA, 2020 |
|
|