Monte Carlo Tree Search as a Tool for Self-Learning and Teaching People to Play Complete Information Board Games

Round 1

Reviewer 1 Report

The authors propose a tool which presents and explains recommendations for playing board games (i.e. Dots and Boxes) generated by a Monte Carlo Tree Search algorithm combined with Neural Networks. Specifically, the manuscript focuses on the UI of the DSS presented to the users, and investigates the learning process resulting from the usage of the proposed system.

In my opinion, the paper needs major revision before possible publication in MDPI Electronics. Specifically, the technical content of the paper should be significantly improved. Detailed comments are provided in the following.

1) The authors should carefully proofread the manuscript to remove grammatical errors and typos.

2) The authors should introduce all the acronyms before using them.

3) The Background section (Sec. II) is extremely dense. The authors should add a Table to summarize the most interesting aspects of each related work grouped by topic.

4) The authors should take into account XAI, transfer learning, and Machine Learning-based approaches used in in state-of-the-art related fields:
[R1] https://doi.org/10.1016/j.jnca.2021.102985
[R2] https://doi.org/10.1016/j.compeleceng.2018.03.005
[R3] https://doi.org/10.1109/TNSM.2021.3098157

5) The authors should clearly point out (e.g., via a bullet list) which are the precise contributions with respect to their previous work (doi:10.3390/app11052056).

6) The description of percentage and visits reported in Fig. 2 is not accurate.

7) In Tab. 3, given that the number of boxes is 25, one of the Avg columns (i.e. Rival or Users) is redundant.

8) The authors claim to employ XAI techniques. However, neither common integrated nor post-hoc techniques are used in the paper. Therefore, in my opinion, the authors are slightly overselling their contribution.

9) The technical contribution of the paper is weak. Indeed, the authors should formalize not only the UI of the proposed DSS but also the Monte Carlo Tree Search (MCTS) algorithm lying in the backend. Also, they claim as a contribution such an MCTS method also in the title of the manuscript.

10) Secs. 6 and 7 are extremely redundant since they express with different words the same concepts. Indeed, in my opinion, in the Discussion section, the authors should focus on limitations and possible improvements of their proposal from both the UI and the backend (i.e. the MCTS algorithm) point of view.

Author Response

Dear Reviewer,

Thank you for the opportunity to revise our paper. We appreciate the careful review and constructive suggestions. It is our believe that the manuscript is substantially improved after making the suggested edits.

Below I attach the document with the reviewer's comments and our responses.

Best regards,
Dr. Edward Núñez

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear Editor, thank you for the opportunity of reviewing the manuscript entitled Monte Carlo Tree Search as a Tool for Self-Learning and Teaching People to Play Complete Information Board Games. The research aims at developing a tool that will present the information in an easily interpretable way and effectively transfer knowledge .

I advice the authors to ask a native speaker to proofread their manuscript.

Furthermore, I have the following suggestions:

-line 17, after "Decision support systems" enter the acronym DSS. From this point on, it is safe to use it. Similarly, the acronym AI is used in line 43 but was not defined. As a general rule, the acronym is defined where the term first appear. Please check the remainder of the document and do the appropriate change. 

-line 20 check the verb "focus in", mostly is used as "focus on" please check here https://dictionary.cambridge.org/dictionary/english/focus and change if appropriate  

-line 111 define SWQL

-specify the source of Table 1 (unless was created by the authors)

The main issues that I raise are:

-the topic of the manuscript which is not correlated with the Journal (Electronics). The authors should argue

-the reference list is not update: only 10 out of the 61 references are from 2020 and 2021. There is an abundancy of literature and I ask the authors to refresh the literature review , accordingly.

 

Author Response

Dear Reviewer,

Thank you for the opportunity to revise our paper. We appreciate the careful review and constructive suggestions. It is our believe that the manuscript is substantially improved after making the suggested edits.

Below I attach the document with the reviewer's comments and our responses.

Best regards,
Dr. Edward Núñez

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors propose a tool which presents and explains recommendations for playing board games (i.e. Dots and Boxes) generated by a Monte Carlo Tree Search algorithm combined with Neural Networks. Specifically, the manuscript focuses on the UI of the DSS presented to the users, and investigates the learning process resulting from the usage of the proposed system.

In my opinion, the paper needs only minor revision before possible publication in MDPI Electronics. Specifically, the authors should carefully proofread the manuscript to remove grammatical errors and typos in the newly added content.

Author Response

Dear reviewer,

Thank you for your advice. We have proofread the entire document, removed grammatical errors and typos, and rephrased some sentences.

Best regards,
Dr. Edward Núñez

Reviewer 2 Report

I have received the new version of the paper and I noticed that the authors applied all the constructive reviews they received. As a result, the new version of the manuscript is improved and I consider that can be considered for being published within the SI of the Electronics.

Author Response

Dear reviewer,

Thank you very much for all your comments.

Best regards,
Dr. Edward Núñez