RI2AP: Robust and Interpretable 2D Anomaly Prediction in Assembly Pipelines

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper introduces a method called Robust and Interpretable 2D Anomaly Prediction (RI2AP) designed to address two dimensions of anomaly prediction, namely when anomalies occur and how they are related. RI2AP uses regression objective for problem modeling to achieve robustness to data rarity. It further includes combining rules as an interpretation mechanism based on a causal-influence framework to provide domain experts with insights into sensor readings and their impact on predictions.  Its performance is evaluated by the Future Factories dataset originating from a manufacturing assembly line designed for rocket assembly. The evaluated performance is compared with that of other related methods. The proposed RI2AP method is shown to outperform related methods, namely the LSTM, Transformer, TimeGPT methods.

The research described in this paper is well motivated and the proposed RI2AP method has advantages. First, RI2AP has 30% improvement in F1 score compared to current ML methods. Second, it tackles the interpretation challenge in anomaly prediction within industrial assembly pipelines. The paper has the merits to be published in the Sensors journal. However, the descriptions of the proposed method are neither straightforward nor explicit. It is thus suggested to describe the proposed method more straightforwardly before the publication. 

Author Response

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions that have been typed in red color in the re-submitted files.

Reviewer Comment 1: "However, the descriptions of the proposed method are neither straightforward nor explicit. It is thus suggested to describe the proposed method more straightforwardly before the publication. "

Response 1:  We have provided clear descriptions for the proposed method in the revised version. The descriptions are included in red color in Section 4- Problem Formulation and in Section 5- The RI2AP Method. 

Reviewer 2 Report

Comments and Suggestions for Authors

1. Provide a clear connection between the Summary, Introduction, Analysis of references, Presentation of the model, results and their analysis and Conclusions with the Objective of the manuscript.

2. Edit references to have their own DOI number, or web page (for example reference 8, etc.).

3. Please, once again check English language.

 

 

Comments on the Quality of English Language

Only, once again check English language.

Author Response

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions that have been typed in red color in the re-submitted files.

Reviewer Comment 1: "Provide a clear connection between the Summary, Introduction, Analysis of references, Presentation of the model, results and their analysis and Conclusions with the Objective of the manuscript. "

Response 1:  We have added a new paragraph that clearly explain the connection at the end of the Introduction Section.

Reviewer Comment 2: "Edit references to have their own DOI number, or web page (for example reference 8, etc.)."

Response 2:  Added the corresponding DOI number for all the references.

 

Reviewer Comment 3: "Please, once again check English language."

Response 3: Checked and corrected the mistakes.

 

 

Reviewer 3 Report

Comments and Suggestions for Authors

The article is well-written and structured

I recommend extending the description of the equations. In addition, the figures that include equations might be improved because the equations are not clear in some of those figures

In my opinion, the article presents a novel contribution.

Author Response

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions that have been typed in red color in the re-submitted files.

Reviewer Comment 1: "The article is well-written and structured"

Response 1:  Thank you so much for the feedback.

Reviewer Comment 2: "I recommend extending the description of the equations. In addition, the figures that include equations might be improved because the equations are not clear in some of those figures"

Response 2:  Expanded the descriptions of the equations and some notations were added for Figure 2.

 

Reviewer Comment 3: "In my opinion, the article presents a novel contribution."

Response 3:  Thank you so much for the feedback.