The (Un)reasonable Effectiveness of Neural Network in Cherenkov Calorimetry^†

Round 1

Reviewer 1 Report

The authors have presented an interesting idea for a novel analysis technique of Cherenkov calorimeter signals using neural networks and complemented the study with a proposal for a longitudinal fiber readout calorimeter in which the z-coordinate is reconstructed from the Cherenkov light time-of-propagation.

The paper is well written and fits well into this special issue and into the journals scope.

I recommend to publish after some minor corrections:

The abstract refers to "plans for a dedicated beam test to validate these findings with a fast, radiation-hard hadron calorimeter prototype". These plans are not discussed in the paper.

Fig. 3 caption: "in he 200" -> "in the 200". The figure layout and the caption are confusing: there is no explanation of 2D-color histogram and there is no (a) nor (b) in the figure.

Fig. 5 caption: " tthe simple sum of Chrenkov" -> " the simple sum of Cherenkov"

Table 1: sigma -> $\sigma$

In the conclusions section it would be good to note if the authors consider their finding of a superior effective neural network technique in Cherenkov calorimetry to be robust against typical nuisance effects in real-world experiments, e.g. timing and gain variations and the SiPMs, pulse shape variations in the fibers, etc. 

The reference section would benefit a lot from hyperlinks to the articles and/or DOI citations. The format is not perfect, between title and journal a comma or any other separator is missing for the first four references.  

Ref. 2 is from year 2021 not 2107.

Ref. 6 should read Hadronic Physics not Phusics

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The term "dual readout" which is used as a reference for comparison to the NN technique is not defined. Unless one grabs Ref. [3] it is not quite clear what "dual" means.  (A sentence or two from the Abstract of Ref. [3] would be sufficient to clarify the matter.)

The geometry specifications are a bit confusing to me: what is the relation between the setup described in lines 40-47 and the one mentioned in lines 94-95 where fibers suddenly appear while previously only layers/plates were mentioned.

For a non-specialist, it might not be clear that e+ are generated from pi0->2g (line 60).  Of course they come from pair production following the gamma emission, but why not say it?

Please define EdgeConv (line 89). Is this an entity of the NN?

The response in Fig. 5 appears to be constant, not linear --- unless "response" is defined in a specific manner.

What is the origin of the star-like structure in the left panel of Fig. 7? I do not see a good physical reason for this kind of anisotropy.

Fix the typos (spell-checking, unmatching singular/plural etc.).

Author Response

Please see the attachment.

Author Response File: Author Response.pdf