SciPost Submission Page
High Efficiency Configuration Space Sampling -- probing the distribution of available states
by Paweł T. Jochym, Jan Łażewski
This is not the latest submitted version.
This Submission thread is now published as
|As Contributors:||Paweł Jochym · Jan Łażewski|
|Date submitted:||2021-04-26 20:27|
|Submitted by:||Jochym, Paweł|
|Submitted to:||SciPost Physics|
Substantial acceleration of research and more efficient utilization of resources can be achieved in modelling investigated phenomena by identifying the limits of system's accessible states instead of tracing the trajectory of its evolution. The proposed strategy uses the Metropolis-Hastings Monte-Carlo sampling of the configuration space probability distribution coupled with physically-motivated prior probability distribution. We demonstrate this general idea by presenting a high performance method of generating configurations for lattice dynamics and other computational solid state physics calculations corresponding to non-zero temperatures. In contrast to the methods based on molecular dynamics, where only a small fraction of obtained data is used, the proposed scheme is distinguished by a considerably higher, reaching even 80%, acceptance ratio and much lower amount of computation required to obtain adequate sampling of the system in thermal equilibrium at non-zero temperature.
Author comments upon resubmission
Your reference: scipost_202101_00011v1
Corresponding author: Paweł T. Jochym Address: Institute of Nuclear Physics, Radzikowskiego 152, 31-342 Cracow, Poland email: email@example.com
Title: High Efficiency Configuration Space Sampling - probing the distribution of available states
Authors: Paweł T. Jochym and Jan Łażewski Type: regular article
Dear Dr Attaccalite,
Thank you for arranging the review of our paper. We are glad that criticism of both referees contributed to the improvement of quality of our work. In fact, both referees noticed strong aspects of our idea and highlighted weak points of the presentation causing possible confusion of the reader. None of the key assumptions of the method have been questioned. The second referee went even further and in his summary rating recognized our approach as of the "high validity and significance" and "top originality".
We thank both referees for careful reading of our text and their valuable remarks. We regret a few mistakes and some deficiencies in presentation pointed by the referees. Following their advice we have made substantial revision of the text and figures correcting all mistakes and omissions as well as extending the explanations to make the presentation clearer.
Please find included a detailed response to both referees. We have already submitted an early response to the first review. This initial response is still valid, but we have reformulated it to closely reflect changes we have made in the text.
We have addressed all points raised by the referees, applied all their suggestions and answered all questions. We believe that after these corrections the manuscript is ready for publication in SciPost Physics without delay.
This resubmission includes: - the revised text, - the detailed reply to both referees (submitted as reply on the submission page), - list of changes
Sincerely Yours, Paweł T. Jochym Jan Łażewski
List of changes
Summary of changes in scipost_202101_00011v1/Jochym&Lazewski:
In response to the referee reports we made the following changes
in the manuscript:
* The Eq. 5 has been corrected.
* The definition of the method was supplemented and the scope of its applicability was more precisely specified.
* Calculations are extended up to 2000 K.
* Figures 3, 4, 5 and 6 were split to several panels to separate data for different temperatures and increase readability of the contents.
* Imprecise statements in our presentation (like "too wild", "very quick", "hardly visible" etc.) were quantified.
* We have modified the text as recommended by the referees - these changes are described in the replies.
* Minor modifications of the text were made in a few other places in order to improve the reading and to remove some typographical errors.
Submission & Refereeing History
You are currently on this page
- Report 2 submitted on 2021-05-19 02:18 by Anonymous
- Report 1 submitted on 2021-05-12 16:31 by Dr Wehinger
Reports on this Submission
Anonymous Report 2 on 2021-5-19 (Invited Report)
- Cite as: Anonymous, Report on arXiv:scipost_202101_00011v2, delivered 2021-05-19, doi: 10.21468/SciPost.Report.2938
The authors of the manuscript "High Efficiency Configuration Space Sampling – probing the distribution of available states" have addressed all the points suggested in my initial report.
The scope and current limitations of the HECSS scheme are now clearly presented. All the previously missing details required for reproducibility are now reported in the manuscript.
I appreciate that Fig. 5 and 6 show the increasing loss of accuracy of the current version of the HECSS as the temperature is increased.
I agree with the authors that high-temperature properties are challenging also using existing MD schemes.
Even if the CLT holds, i.e., even if the variance is finite, the convergence to the expected normal distribution can get rather slow.
Perhaps this can be better understood in terms of the tails of the potential energy distribution --- not clearly visible in Fig. 2 --- which get "heavier" at high-temperature because of the anharmonicity.
Investigating the role of these tails, e.g., perturbatively, may reveal a strategy to improve the agreement between the HECSS and MD without resorting to a somehow arbitrary "temperature calibration".
Report 1 by Bjorn Wehinger on 2021-5-12 (Invited Report)
- Cite as: Bjorn Wehinger, Report on arXiv:scipost_202101_00011v2, delivered 2021-05-12, doi: 10.21468/SciPost.Report.2910
Referee report for the revised manuscript entitled "High Efficiency Configuration Space Sampling – probing the distribution of available states".
The authors have carefully addressed all my comments in the revised version of the manuscript. The new version of the manuscript properly describes the new method and nicely illustrates its application and limitations to lattice dynamics calculations at finite temperatures.
As I have mentioned in my first report, the idea is highly original and clearly opens a new pathway in the study of thermal properties at finite temperatures applicable to large systems that are difficult to address otherwise.
The revised version of the manuscript fulfills all general acceptance criteria. I believe that it is of great interest to a broad audience and thus recommend publication in SciPost Physics.
I only have a few minor comments:
1. How does the applied temperature adjustment of 90K mentioned for the calculation of phonon life times at a temperature of 600K compare to an estimated correction obtained by "deriving the $C_n$ coefficients from the force constants matrices" as explained in Section 2?
2. Labels in Fig. 5 disagree with description in the figure caption.