Hello reviewers of NAI journal submission 671-1651:

Thank you for your encouraging feedback with respect to our original submission. And thank you for your helpful suggestions for how to strengthen the paper.  We have responded to all of them. We like the result and we hope you do too.

This document:
* summarises your suggestions for improvements and points to where you can find our responses in our 1st revision
* explains the colour coding present in the 1st revision
* makes some closing remarks.



Reviewer 1 suggestions:

R1A: say more about the limitations of OWL embeddings, for example how the neighbourhood-based or fixed-length random walk-based embeddings lose potential long-range logical relationships
* our response is in Section 2.3, paragraph 7

R1B: say more about the approaches for knowledge injection in Section 2.3
* our response is in Section 2.3, paragraphs 4, 5 and 6

R1C: consider discussing the topic of translating OWL to FOL in Section 3.4
* our response is in Section 3.4, in the final two paragraphs



Reviewer 2 suggestions:

R2A: expand the overview of neurosymbolic AI to better contextualise the role ontology reasoning can play
* our response is in Section 1 (the Introduction), in paragraph 3

R2B: add more concrete examples demonstrating OWL reasoning in action to reinforce the arguments made, the main focus seems to be Visual Relationship Detection
* we now describe and cite several (6) of (the limited number of) examples of OWL reasoning in action we were able to find in the literature 
* we have placed these discussions where we feel they best sit, so our response here is distributed
* see Section 2.4, paragraphs 5 and 7 
* see Section 3.1, paragraph 1



Colour coding:

We introduced some colour coding to the document to highlight substantive changes.  We did this for ourselves, initially, but decided you might find it useful, so we retained it in the pdf we submitted.  Paragraphs coloured blue have either been introduced in response to your suggestions for improvements or have been substantially reworked in the process of preparing the 1st revision. The reworkings have mostly to do with polishing language so the paper flows better. But there are one or two instances where bits of new content have crept in as well.  Several smaller, surgical insertions of text are coloured blue as well.  Everything not coloured blue is either identical to the original submission or has been adjusted at most only superficially. 



Additional points:

The original submission talked of OWL-based KGs as having 'agency' and as being capable of being used as 'active agents'. We have dropped this language and introduced language around being able to use OWL-based KGs as 'symbolic deduction engines'. We feel this new language better conveys our intended meaning.

The size of the paper has increased with this 1st revision --- from 10 to 11.5 pages. An increase of up to 2 pages was authorised by the journal's executive editors (Pascal) via email, and this is what has allowed us to respond to your suggestions for more content as thoroughly as we have. Please do not be alarmed by the increase in size.

Finally, we apologise for being slow to submit this 1st revision. We hope we have not put yourselves in awkward situations as a result.



Thank you and kind regards,

David Herron