Enhancing Inference on Physiological and Kinematic Periodic Signals via Phase-Based Interpretability and Multi-Task Learning
Round 1
Reviewer 1 Report
The paper would be more orderly with a section of conclusions, which in this case is what the authors have called discussion. The results themselves should be separated from the discussion. It would also be interesting to emphasize the usefulness of this model by mentioning real cases of its application. The reference section appears in the middle of the annex.
Author Response
Dear editor,
We are re-submitting a revised version of the manuscript “Enhancing Inference on Physiological and Kinematic Periodic Signals via Interpretability and Multi-task Learning”. The reviewer’s concerns and comments were taken into consideration and addressed by appropriate changes in the manuscript. We greatly appreciate the reviewer’s efforts since their feedback helped us improve our manuscript significantly. Our changes include a revised introduction and methodology to clarify our contributions and scope, more details on the experimental setup. The complete list of changes can be viewed in the attached PDF document that contains our responses to the reviewers’ comments.
Sincerely yours,
Reza Soleimani (on behalf of the Authors)
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The authors propose an approach to development of an interpretable model for periodic signals. It is an interesting topic however the description needs improvement to be understandable for readers.
The authors should explain what they mean by interpretability.
Some steps in the processing need clarification. For example, page 6, line 134: it is stated that "the hyper-parameters lambda1 and lambda2 arechonse empirically". Please explain what the range of values is and which values are the best. And how did you arrive at these values?
Page 8, description on lines 191-195 is not clear.
Page 4, line 203: what are the values of lambda1, lambda2 and lambda3?
Page 4, line 205: AdamW and RMSprop - missing references
Page 4, line 210: "the best lambda1 and lambda2" - what are the best values?
Learning is not clearly described. It must be described in more detail in relation to the solved task.
It is not clear from the description how to arrive at the correct interpretation of the results. It needs more detailed explanation and a practical example.
English needs very thorough revision. There are many grammatical errors and complicated formulations that make reading difficult.
Author Response
Dear editor,
We are re-submitting a revised version of the manuscript “Enhancing Inference on Physiological and Kinematic Periodic Signals via Interpretability and Multi-task Learning”. The reviewer’s concerns and comments were taken into consideration and addressed by appropriate changes in the manuscript. We greatly appreciate the reviewer’s efforts since their feedback helped us improve our manuscript significantly. Our changes include a revised introduction and methodology to clarify our contributions and scope, more details on the experimental setup. The complete list of changes can be viewed in the attached PDF document that contains our responses to the reviewers’ comments.
Sincerely yours,
Reza Soleimani (on behalf of the Authors)
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The manuscript has been significantly improved. However, there are still some unclear points.
There are also some imprecise formulations.
Page 6: it is not clear how the middle phase values are applied. PLease explain in more detail.
Page 8: more details on z-standardization and augmentation are needed.
Figure 7: I have still some concerns about the modelling and reconstruction tasks in ECG. THe QRS complex is generated and predicted well. However, the other waves (P, T) are not very close to the true signal. How would the authors solve this problem in case of pathological signals where these waves have their importance in diagnostics?
The language still needs revision. There are some grammatical and mistyping errors.
Page 7, paragraph after line 162: there is "Fig. ??" - please add the right reference.
Author Response
Dear editor,
We are re-submitting a revised version of the manuscript “Enhancing Inference on Physiological and Kinematic Periodic Signals via Interpretability and Multi-task Learning”. The reviewer’s concerns and comments were taken into consideration and addressed by appropriate changes in the manuscript. We greatly appreciate the reviewer’s efforts since their feedback helped us improve our manuscript even further compared to the first revision. The changes can be seen in their corresponding sections. The complete list of changes can be viewed in the attached PDF document that contains our responses to the reviewers’ comments.
Sincerely yours,
Reza Soleimani (on behalf of the Authors)
Author Response File: Author Response.pdf
