Analysis of Cross-Referencing Artificial Intelligence Topics Based on Sentence Modeling

Round 1

Reviewer 1 Report

Congratulations. I  recommend to publish this article after  minor corrections.

Comments for author File: Comments.pdf

Author Response

1. The title of the paper reflects the content of the paper, (please describe, why you use cross referencing in the title, it can be for example: Analysis of Artificial Intelligence Topics based on Sentence Modeling).

Response)
This study focused on how diverse a topic relates to a range of knowledge, rather than on the purpose of analyzing each subject. In other words, the perspective of cross-referencing was used to identify which areas of knowledge related to AI are comprehensively related to the CS field. Cross-referencing of topics was used to indicate that AI-related topics are related to various knowledge areas of CS, rather than one knowledge area.

2. Line 37-46 present interesting references, Have you any examples from Europe?

Response)
Various AI-related efforts are also underway in Europe. For example, in January 2018, the Danish government announced a digital growth strategy focused on AI. The United Kingdom also announced the importance of establishing rules and standards for AI technology and AI at the World Economic Forum held in January 2018. In addition, Austria's Ministry of Infrastructure Resources established the Robots Council in August 2017 to conduct research on AI-related business requirements, legal requirements, and social and ethical values.
In order for such diverse policies to be carried out through education, they must be set out in a “curriculum” document. However, despite various policy announcements, there are no official documents yet for the countries cited above. Therefore, the above cases were not presented in the paper.

3. It is worth underline that authors based their methods on literature - CNN, Manhattan LSTM, and multi-head attention networks used experimental procedure. Nowadays, there is no research available on the curriculum relevance of the topics.

Response)
This study analyzed topics related to AI that should be considered for education related to AI from the perspective of CS terminology. The more data there are, the more stable the topic modeling can produce. However, due to the nature of this study, there are limitations because it is based on curriculum documents that focus only on key topics related to AI education. It is similar to the limitations of recent studies such as Sekiya (2014) and Yoshitatsu (2017). Therefore, in the future, if documents are presented in more countries, we would like to conduct further analyses using those documents. We have outlined these limitations in the revised document.

4. Lines 126-128 - Figure 1. Structure of CNN and 202-204 lines Figure 4 please make more readable (the words in a bigger size).

Response)
We have revised the figures and font sizes to provide more clarity in accordance with your suggestion.

5. In the section 2. Related Research, 137-143 lines, where is described LSTM . Try to prepare more clear.

Response)
Line 133-148.
Based on the reviewer's opinion, We have supplemented the contents as follows.
“RNNs process word inputs in a particular order and learn from the order of appearance of particular expressions. Thus, modeling can process semantic similarities among sentences and phrases. RNNs can also use their own models, but they are used as an extension of the modified model. Manhattan long short-term memory (LSTM), which is a modified RNN model, uses two LSTMs to read and process word vectors that represent two input sentences. This model has been shown to outperform the complex neural network model. The structure of the Manhattan LSTM is shown in Figure 2 (Mueller and Thyagarajan 2016). The main feature of this model is that LSTM and the Manhattan metric are used based on a Siemens networks structure that contains two identical subnetwork components. The hidden state h1 is learned through word vectors and randomly generated weights. Then, the hidden state sentence ht is generated based on the input function using hidden state ht-1 and position t. Subsequently, the semantic similarity of sentences is measured using the vectors of the final hidden state. For example, “He,” “is,” and “smart” are the words in the vectors xi; x1 is the input vector of h1 and is used to calculate the status value of h1; h2 is calculated by referring to the previous state value and x2; and the final hidden state, h3, is calculated through x3 and the previous state value h2. “A,” “truly,” “wise,” and “man” are treated in the same manner, and similarities are measured with the final vectors calculated by processing two sentences.

6. What's the software was being used for calculation of all obtained results?

Response)
Line. 186-187.
This system was developed in Python 3.6 and executed on Linux 16.04.
We have added that information.

7. I recommend to indicate the limitations of the method and characterize the directions of the future research.

Response)
Line492-500.
We have added the following information as regards the limitations of this study.

“Among the limitations of this study is sparsity of data on the core topics used in education. A large amount of data is needed to extract stable values through machine learning, but the ultimate limitation is that documents on the curriculum or core topics of education are not sufficient in terms of learning data. Furthermore, it is an area where research has not been conducted sufficiently on how to justify the results after the study through learning models. Consequently, identifying the possibilities through the educational use of the research results is time-consuming. To reflect emerging knowledge in education, the composition of knowledge or modeling of topics is essential. To achieve this, it is necessary to establish various methodologies to overcome limitations in the availability of extracted knowledge.”

8. The presentation of literature material is complete and logical. List of references has of the contemporary sources (not older than 5 years) and the international researches. I think it's a very good result, but it is worth to add some references from previous papers of this journal.

Response)
In accordance with your suggestion, we have added two new references from this journal.

9. The conclusions are accurate and supported by the content.
Response)
Thank you for your helpful comments on the research.

Author Response File: Author Response.docx

Reviewer 2 Report

I read the article with real pleasure. I think that it is prepared very correctly and clearly. The research process has been presented in an orderly manner and the reader has the opportunity to learn the entire methodology and how to conduct research.
In addition, I think that the topic chosen by you is extremely interesting and I will be happy to observe further research in your chosen direction

Author Response

I read the article with real pleasure. I think that it is prepared very correctly and clearly. The research process has been presented in an orderly manner and the reader has the opportunity to learn the entire methodology and how to conduct research.
In addition, I think that the topic chosen by you is extremely interesting and I will be happy to observe further research in your chosen direction

 

Response)
Thank you for your opinion.
With this paper, we hope to advance research in this interesting area.

Author Response File: Author Response.docx

Reviewer 3 Report

The methodology of the current scientific study seems well devised, each of its steps are clearly presented. The results seem to support the hypothesis set in the beginning of the manuscript and the findings are well structured in the "Conclusions" section. Regarding the manuscript's scientific soundness, I have no comments.

However, a minor English spellcheck and text editing is required (e.g. Table 6 - the word "Fundamental" mustn't be divided and the columns should be equal in width) is required and most importantly I highly recommend that ALL figures should be replaced with ones having a higher resolution (at least 300 DPI) as they are blurry and can't be fully observed.

Author Response

The methodology of the current scientific study seems well devised, each of its steps are clearly presented. The results seem to support the hypothesis set in the beginning of the manuscript and the findings are well structured in the "Conclusions" section. Regarding the manuscript's scientific soundness, I have no comments.

However, a minor English spellcheck and text editing is required (e.g. Table 6 - the word "Fundamental" mustn't be divided and the columns should be equal in width) is required and most importantly I highly recommend that ALL figures should be replaced with ones having a higher resolution (at least 300 DPI) as they are blurry and can't be fully observed.

 

Response)
Thank you for your suggestions.
We have corrected the table and the spelling of the words.
All figures have also been replaced with others having higher resolutions.

 

Author Response File: Author Response.docx