A Novel Fusion Pruning Algorithm Based on Information Entropy Stratification and IoT Application

Round 1

Reviewer 1 Report

The topic considered by the authors is a hot topic in the context of deep learning. The solution they propose is also interesting. The experiments illustrated are interesting.

However, there is a big problem: the authors do not explain what this paper has to do with the special issue to which they submitted the paper. The only possible connection concerns the use of IoT. However, the term "IoT" is present only in the title of the paper!

The authors should make major revisions in the paper. In particular, they should insert an ad hoc (and very long) section to explain how their approach is related to the IoT world, considering not only the classic IoT architectures, but also the most recent and innovative ones, e.g. Social IoT (see the papers of Iera, Atzori et al.) and Multi-IoT (see the papers of Baldassarre, Virgili, Cauteruccio, Fortino et al).

Author Response

Please see the attachment.

Author Response File: Author Response.doc

Reviewer 2 Report

This paper is about proposing a fusion method for pruning to compact the proposed deep learning model using information entropy stratification and BN layer scaling. factor. The pruning is applied between the pertaining and the fine-tuning phases.  

-please use the full name of BN in the abstract.  

Author Response

Please see the attachment.

Author Response File: Author Response.doc

Round 2

Reviewer 1 Report

I still think this paper has little to do with IoT. However, the authors at least made an effort to add a section to try to convince the reader otherwise. Therefore, at least for the effort made by the authors, I vote for the paper to be published.

Author Response

Please see the attachment.

Author Response File: Author Response.doc