Abstract
Proposed herein are novel techniques that utilize zero-shot learning and superpixels in order to efficiently predict an optimized configuration profile for a given network. The techniques may be able to not only efficiently handle large amounts of data for the prediction, but also account for any dynamic network changes without the need of manual intervention. The techniques may perform well with dynamic unseen and/or out-of-training sample changes. Such techniques may be useful for efficiently creating network configurations that can improve network performance even if the network experiences structural changes.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
S Liang, Eric; Y Chour, Claire; Shao, Qihong; and Zacks, Dave, "ZERO-SHOT SUPERPIXEL LEARNING FOR NETWORK CONFIGURATION OPTIMIZATION", Technical Disclosure Commons, (May 29, 2024)
https://www.tdcommons.org/dpubs_series/7049