A Review on the Application of Machine Learning in Gamma Spectroscopy: Challenges and Opportunities
, Nahid Madani 3, Dariush Sardari 4 and Bashir Bashiri 1,5,*Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe exploration of machine learning (ML) methodologies, evaluation techniques, and verification processes is crucial. This paper presents these concepts very clearly in an excellent review. The paper can be accepted for publication.
One recommendation: The inclusion of concrete examples can significantly enhance the understanding of complex concepts in a paper, especially in fields like machine learning (ML) applied to gamma spectroscopy. By providing one or two figures illustrating real-world applications or results, the authors could offer a clearer application of ML for Gamma spectroscopy.
The link with gamma spectroscopy is missing.
Author Response
Response to Reviewer 1 comments
Dear Reviewer,
Thank you for your valuable comments. We would like to express our sincere appreciation for your input.
We have thoroughly revised the previous version of the manuscript based on your constructive comments. In response to your suggestions, section 2-5 is added to explain the link between ML and gamma spectroscopy. Figure 5 is added also to make it even more understandable for the readers. Also in response to comments from other reviewers, we added table 1 that summarizes the literature reviewed in this study.
Reviewer 2 Report
Comments and Suggestions for AuthorsThe article reviews recent studies on the application of Machine Learning (ML) in gamma spectroscopy. The authors discuss the basic principles of artificial intelligence (AI) and various types of machine learning models. They review real-life case studies where ML models have been used to improve gamma spectroscopy, including radioisotope identification, optimizing detector performance, and simplifying environmental monitoring processes. The article also addresses challenges associated with applying ML methods in gamma spectroscopy, such as computational efficiency, the impact of data normalization, transparency and trust, the need for expertise, and the generalization to unknown conditions.
In my opinion, the authors have done an excellent job of reviewing the previous studies on the application of ML in gamma spectroscopy. The article discusses exciting opportunities to enhance gamma spectroscopy using ML models in areas such as automated identification, optimization of detector performance, the development of directional gamma-ray spectrometry, background correction in low-activity measurements, and more. These discussions could inspire various research groups to explore these ideas, potentially leading to interesting and valuable results.
I think this paper is suitable for Spectroscopy Journal. I would recommend accepting the paper at its current state.
Author Response
Dear Reviewer
Thank you very much for your time and efforts put into reviewing our manuscript.
Best regards.
Reviewer 3 Report
Comments and Suggestions for AuthorsLine 34: Please revise it to “By detecting and analyzing….”
General comment: When you describe literature, you may consider summarizing them in a table where each row is a study and each column is an attribute such as sample type, sample size, preprocessing method, algorithm and key parameters, model performance. This could free you from repetitive description and may help you be more specific on the highlights of each study.
Line 213-232: Could you please be more specific and quantitative if possible on the results and conclusions of these two studies? For example, what’s the classification accuracy?
Line 258-261: Please be more specific and quantitative on the outcome and conclusion.
Line 315-316: Please be more specific on the accuracy and how much longer it takes to predict compared to other algorithms.
Line 318-321: Could you describe the results and conclusion?
Line 339-341: Please be more quantitative in reporting the results.
Section 3-5: Please be more specific on the result and conclusion of each study.
Line 498: Please revise it to “…to evaluate the performance of an ML classification model,…”.
Section 3.6: These metrics are for classification problems. Is there a reason why you didn’t mention any metrics for regression?
Author Response
Dear Reviewer,
Thank you for your valuable comments. We would like to express our sincere appreciation for your input.
We have thoroughly revised the previous version of the manuscript based on your constructive comments. In response to your suggestions, we added table 1 that summarizes the literature reviewed in this study. Also in response to comments from other reviewers, the link between M and gamma spectroscopy is explained explicitly in section 2-5. Figure 5 is added also to make it even more understandable for the readers.
Please find below our point-by-point responses to your comments.
Comment 1: Line 34: Please revise it to “By detecting and analyzing….”
Response: This sentence has been revised accordingly
Comment 2: Line 213-232: Could you please be more specific and quantitative if possible on the results and conclusions of these two studies? For example, what’s the classification accuracy?
Response: This comment has been taken into account and in section 3.1 the accuracy of the studies are added in the text.
Comment 3: Line 258-261: Please be more specific and quantitative on the outcome and conclusion.
Response: The second paragraph of section 3.2 is revised and quantitative results are added in the text.under the section 3
Comment 4: Line 315-316: Please be more specific on the accuracy and how much longer it takes to predict compared to other algorithms.
Response: This comment has been taken into account and the second paragraph of the section 3.3 is revised accordingly and quantitative results are now included in this paragraph.
Comment 5: Line 318-321: Could you describe the results and conclusion?
Response: Paragraph 3 in section 3.3 is expanded considering the suggestion in this comment.
Comment 6: Line 339-341: Please be more quantitative in reporting the results.
Response: This paragraph is revised accordingly.
Comment 7: Section 3-5: Please be more specific on the result and conclusion of each study.
Response: Considering the proposal in this comment, literature review in section 3-5 is revised critically and the quantitative results are added, given that those results were provided in the corresponding article.
Comment 8: Line 498: Please revise it to “…to evaluate the performance of an ML classification model,…”.
Response: This sentence is revised.
Comment 9: Section 3.6: These metrics are for classification problems. Is there a reason why you didn’t mention any metrics for regression?
Response: According to the literature review, classification algorithms are widely used in gamma spectroscopy tasks. Therefore here the metrics for the classification algorithms are explained. This argument is justified now is section 3.6.
Comment 10: General comment: When you describe literature, you may consider summarizing them in a table where each row is a study and each column is an attribute such as sample type, sample size, preprocessing method, algorithm and key parameters, model performance. This could free you from repetitive description and may help you be more specific on the highlights of each study.
Response: In response to this comment the literature review is summarized in table 1.
Reviewer 4 Report
Comments and Suggestions for AuthorsThe manuscript presents a nice machine learning analysis study that could be of interest to many readers; first of all, beginners in the field of applications of machine learning.
Authors shortly narrate many published papers in the field of Gamma spectroscopy. That is the only connection with the gamma spectroscopy, because all described machine learning algorithms are equally applicable to other kinds of spectroscopy. The large part of the paper describing machine learning could be used, also, to write the paper about, for example, laser spectroscopy.
Opportunities and challenges are also very general, applicable for other spectroscopy techniques.
Authors should provide the discussion what is specific, if anything, to the data obtained in Gamma spectroscopy compared to the other spectroscopy data. Or, to point out that ML is also applicable to the Gamma spectroscopy as to other spectroscopy techniques.
Author Response
Dear Reviewer,
Thank you for your valuable comments. We would like to express our sincere appreciation for your input.
We have thoroughly revised the previous version of the manuscript based. In response to your suggestions, section 6 is newly added. In response to other reviewers, section 2-5 is added to explain the link between ML and gamma spectroscopy. Figure 5 is added also to make it even more understandable for the readers. Also, in response to comments from other reviewers, we added table 1 that summarizes the literature reviewed in this study.
Comment: Authors should provide the discussion what is specific, if anything, to the data obtained in Gamma spectroscopy compared to the other spectroscopy data. Or, to point out that ML is also applicable to the Gamma spectroscopy as to other spectroscopy techniques.
Response: In response to this comment a new section (6) is added titled ‘Data Collection for ML in Gamma Spectroscopy and Other Spectroscopic Techniques’ in which this issue has been shortly discussed.
