Neural network analytic continuation for Monte Carlo: Improvement by statistical errors
Kai-Wei Sun(孙恺伟)1 and Fa Wang(王垡)1,2,†
1 International Center for Quantum Materials, School of Physics, Peking University, Beijing 100871, China; 2 Collaborative Innovation Center of Quantum Matter, Beijing 100871, China
Abstract This study explores the use of neural network-based analytic continuation to extract spectra from Monte Carlo data. We apply this technique to both synthetic and Monte Carlo-generated data. The training sets for neural networks are carefully synthesized without "data leakage". We find that the training set should match the input correlation functions in terms of statistical error properties, such as noise level, noise dependence on imaginary time, and imaginary time-displaced correlations. We have developed a systematic method to synthesize such training datasets. Our improved algorithm outperforms the widely used maximum entropy method in highly noisy situations. As an example, our method successfully extracted the dynamic structure factor of the spin-1/2 Heisenberg chain from quantum Monte Carlo simulations.
Fund: FW acknowledges support from the National Natural Science Foundation of China (Grant Nos. 12274004 and 11888101). Quantum Monte Carlo simulations are performed on TianHe-1A of National Supercomputer Center in Tianjin.
Corresponding Authors:
Fa Wang
E-mail: wangfa@pku.edu.cn
Cite this article:
Kai-Wei Sun(孙恺伟) and Fa Wang(王垡) Neural network analytic continuation for Monte Carlo: Improvement by statistical errors 2023 Chin. Phys. B 32 070705
[1] White S, Scalapino D, Sugar R and Bickers N 1989 Phys. Rev. Lett.63 1523 [2] Silver R N, Sivia D S and Gubernatis J E 1990 Phys. Rev. B41 2380 [3] Henelius P, Sandvik A W, Timm C and Girvin S 2000 Phys. Rev. B61 364 [4] Sandvik A W 1998 Phys. Rev. B57 10287 [5] Shao H and Sandvik A W 2023 Phys. Rep.1003 1 [6] Sandvik A W 2016 Phys. Rev. E94 063308 [7] LeCun Y, Bengio Y and Hinton G 2015 Nature521 436 [8] Yoon H, Sim J H and Han M J 2018 Phys. Rev. B98 245101 [9] Arsenault L F, Neuberg R, Hannah L A and Millis A J 2017 Inverse Problems33 115007 [10] Xie X, Bao F, Maier T and Webster C 2021 Analytic continuation of noisy data using adams bashforth residual neural network Tech. rep. Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States) [11] Huang D and Yang Y F 2022 Phys. Rev. B105 075112 [12] Zhang R, Merkel M E, Beck S and Ederer C 2022 Phys. Rev. Research4 043082 [13] Yao J, Wang C, Yao Z and Zhai H 2022 Machine Learning: Science and Technology3 025010 [14] Fournier R, Wang L, Yazyev O V and Wu Q 2020 Phys. Rev. Lett.124 056401 [15] Kaufman S, Rosset S, Perlich C and Stitelman O 2012 ACM Transactions on Knowledge Discovery from Data (TKDD)6 1 [16] Jordan M I and Mitchell T M 2015 Science349 255 [17] Domingos P 2012 Communications of the ACM55 78 [18] Giles C L and Maxwell T 1987 Appl. Opt.26 4972 [19] Novak R, Bahri Y, Abolafia D A, Pennington J and Sohl-Dickstein J 2018 arXiv preprint arXiv:1802.08760 [20] Albawi S, Mohammed T A and Al-Zawi S 2017 International conference on engineering and technology (ICET) (IEEE) pp. 1-6 [21] He K, Zhang X, Ren S and Sun J 2016 Deep residual learning for image recognition Proceedings of the IEEE conference on computer vision and pattern recognition pp. 770-778 [22] Yu F,Wang D, Shelhamer E and Darrell T 2018 Deep layer aggregation Proceedings of the IEEE conference on computer vision and pattern recognition p. 2403-2412 [23] Ramachandran P, Zoph B and Le Q V 2017 arXiv preprint arXiv:1710.05941 [24] Agarap A F 2018 arXiv preprint arXiv:1803.08375 [25] Iosifidis A and Tefas A 2022 Deep learning for robot perception and cognition [26] Srivastava N, Hinton G, Krizhevsky A, Sutskever I and Salakhutdinov R 2014 The journal of machine learning research15 1929 [27] Joyce J M 2011 Kullback-leibler divergence International encyclopedia of statistical science (Springer) pp. 720-722 [28] Chollet F, et al. 2015 Keras https://keras.io [29] Abadi M, et al. 2015 Tensor Flow: Large-scale machine learning on heterogeneous systems software [30] Kingma D P and Ba J 2014 arXiv preprint arXiv:1412.6980 [31] Kraberger G J, Triebl R, Zingl M and Aichhorn M 2017 Phys. Rev. B96 155128 [32] Bergeron D and Tremblay A M S 2016 Phys. Rev. E94 023303 [33] Altmann A, Toloşi L, Sander O and Lengauer T 2010 Bioinformatics26 1340 [34] Sandvik A W 1999 Phys. Rev. B59 R14157 [35] Okamoto S, Alvarez G, Dagotto E and Tohyama T 2018 Phys. Rev. E97 043308 [36] Pan S J and Yang Q 2010 IEEE Transactions on knowledge and data engineering22 1345
Altmetric calculates a score based on the online attention an article receives. Each coloured thread in the circle represents a different type of online attention. The number in the centre is the Altmetric score. Social media and mainstream news media are the main sources that calculate the score. Reference managers such as Mendeley are also tracked but do not contribute to the score. Older articles often score higher because they have had more time to get noticed. To account for this, Altmetric has included the context data for other articles of a similar age.