Distributed learning is a relatively young area as compared to (parametric) decentralized detection and estimation, wireless sensor networks (WSNs), and machine learning. This paper decomposes the literature on distributed learning according to two general research themes: distributed learning in WSNs with a fusion center, where the focus is on how learning is effected when communication constraints limit access to training data; and distributed learning in WSNs with in-network processing, where the focus is on how intersensor communications and local processing may be exploited to enable communication-efficient collaborative learning. Both themes are discussed within the context of several papers in the field.
All Science Journal Classification (ASJC) codes
- Signal Processing
- Electrical and Electronic Engineering
- Applied Mathematics