Hello veterans!
I am interested in working on the intersection of massive Machine Type Communication (mMTC) and Machine Learning. Specifically, I am trying to address the challenges where integration of ML can provide potential solutions. Thanks in advance!
Indeed it is a big challenge. In particular, when working Bayesian parametric framework of maximum likelihood estimation, we face the normalizing constant, i.e., how to integrate it exactly? The exact calculation, is very difficult, we therefor recourse to some method of approximation, e.g., Laplace Approximation for calculating very high dimensional integrals as well as Laplace approximation. Asymptotic series approximation is also one way, but it is still not general. The real need is to evaluate closed form or general solution or analytic solution???
Hi Shah, this topic is really interesting. MMTC is a key characteristic of 5G communication services to scale the connectivity demand for the expanding number of wireless network-enabled devices, focusing on the efficient transmission of small payloads over an extended coverage area. Emerging applications such as body-area networks, smart homes, internet of things (IoT), and drone delivery will generate sporadic traffic between a massive number of geographically spread equipment, which will generate the data you need to apply machine learning.
Check the following article I wrote recently, especially section III-B, where you will find the potential of machine learning to support mMTC.
Article Machine Learning for 5G/B5G Mobile and Wireless Communicatio...
Hope this can help!
- Badges
- Science topic