Automated Continual Learning of Defect Identification in Coherent Diffraction Imaging

Orcun Yildiz; Henry Chan; Krishnan Raghavan; William Judge; Mathew J. Cherukara; Prasanna Balaprakash; Subramanian Sankaranarayanan; Tom Peterka

doi:10.1109/AI4S56813.2022.00006

Automated Continual Learning of Defect Identification in Coherent Diffraction Imaging

Orcun Yildiz, Henry Chan, Krishnan Raghavan, William Judge, Mathew J. Cherukara, Prasanna Balaprakash, Subramanian Sankaranarayanan, Tom Peterka

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

2 Scopus citations

Abstract

X-ray Bragg coherent diffraction imaging (BCDI) is widely used for materials characterization. However, obtaining X-ray diffraction data is difficult and computationally intensive. Here, we introduce a machine learning approach to identify crystalline line defects in samples from the raw coherent diffraction data. To automate this process, we compose a workflow coupling coherent diffraction data generation with training and inference of deep neural network defect classifiers. In particular, we adopt a continual learning approach, where we generate training and inference data as needed based on the accuracy of the defect classifier instead of all training data generated a priori. The results show that our approach improves the accuracy of defect classifiers while using much fewer samples of data.

Original language	English
Title of host publication	Proceedings of AI4S 2022
Subtitle of host publication	Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	1-6
Number of pages	6
ISBN (Electronic)	9781665462075
DOIs	https://doi.org/10.1109/AI4S56813.2022.00006
State	Published - 2022
Externally published	Yes
Event	3rd IEEE/ACM International Workshop on Artificial Intelligence and Machine Learning for Scientific Applications, AI4S 2022 - Dallas, United States Duration: Nov 13 2022 → Nov 18 2022

Publication series

Name	Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis

Conference

Conference	3rd IEEE/ACM International Workshop on Artificial Intelligence and Machine Learning for Scientific Applications, AI4S 2022
Country/Territory	United States
City	Dallas
Period	11/13/22 → 11/18/22

Funding

This material is based upon work supported by the U.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing Research, under contract number DE-AC02-06CH11357. We gratefully acknowledge the computing resources provided on Swing, a high-performance computing cluster operated by the Laboratory Computing Resource Center at Argonne National Laboratory. Work performed at the Center for Nanoscale Materials and Advanced Photon Source, both U.S. Department of Energy Office of Science User Facilities, was supported by the U.S. DOE, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357. This material is based upon work supported by the U.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing Research, under contract number DEAC02-06CH11357. We gratefully acknowledge the computing resources provided on Swing, a high-performance computing cluster operated by the Laboratory Computing Resource Center at Argonne National Laboratory. Work performed at the Center for Nanoscale Materials and Advanced Photon Source, both U.S. Department of Energy Office of Science User Facilities, was supported by the U.S. DOE, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357.

Keywords

HPC workflows
catastrophic forgetting
continual learning
defect identification

Access to Document

10.1109/AI4S56813.2022.00006

Cite this

Yildiz, O., Chan, H., Raghavan, K., Judge, W., Cherukara, M. J., Balaprakash, P., Sankaranarayanan, S., & Peterka, T. (2022). Automated Continual Learning of Defect Identification in Coherent Diffraction Imaging. In Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis (pp. 1-6). (Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/AI4S56813.2022.00006

Yildiz, Orcun ; Chan, Henry ; Raghavan, Krishnan et al. / Automated Continual Learning of Defect Identification in Coherent Diffraction Imaging. Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis. Institute of Electrical and Electronics Engineers Inc., 2022. pp. 1-6 (Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis).

@inproceedings{95855b055d294309827990e85124b4a2,

title = "Automated Continual Learning of Defect Identification in Coherent Diffraction Imaging",

abstract = "X-ray Bragg coherent diffraction imaging (BCDI) is widely used for materials characterization. However, obtaining X-ray diffraction data is difficult and computationally intensive. Here, we introduce a machine learning approach to identify crystalline line defects in samples from the raw coherent diffraction data. To automate this process, we compose a workflow coupling coherent diffraction data generation with training and inference of deep neural network defect classifiers. In particular, we adopt a continual learning approach, where we generate training and inference data as needed based on the accuracy of the defect classifier instead of all training data generated a priori. The results show that our approach improves the accuracy of defect classifiers while using much fewer samples of data.",

keywords = "HPC workflows, catastrophic forgetting, continual learning, defect identification",

author = "Orcun Yildiz and Henry Chan and Krishnan Raghavan and William Judge and Cherukara, {Mathew J.} and Prasanna Balaprakash and Subramanian Sankaranarayanan and Tom Peterka",

note = "Publisher Copyright: {\textcopyright} 2022 IEEE.; 3rd IEEE/ACM International Workshop on Artificial Intelligence and Machine Learning for Scientific Applications, AI4S 2022 ; Conference date: 13-11-2022 Through 18-11-2022",

year = "2022",

doi = "10.1109/AI4S56813.2022.00006",

language = "English",

series = "Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "1--6",

booktitle = "Proceedings of AI4S 2022",

}

Yildiz, O, Chan, H, Raghavan, K, Judge, W, Cherukara, MJ, Balaprakash, P, Sankaranarayanan, S & Peterka, T 2022, Automated Continual Learning of Defect Identification in Coherent Diffraction Imaging. in Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis. Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis, Institute of Electrical and Electronics Engineers Inc., pp. 1-6, 3rd IEEE/ACM International Workshop on Artificial Intelligence and Machine Learning for Scientific Applications, AI4S 2022, Dallas, United States, 11/13/22. https://doi.org/10.1109/AI4S56813.2022.00006

Automated Continual Learning of Defect Identification in Coherent Diffraction Imaging. / Yildiz, Orcun; Chan, Henry; Raghavan, Krishnan et al.
Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis. Institute of Electrical and Electronics Engineers Inc., 2022. p. 1-6 (Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Automated Continual Learning of Defect Identification in Coherent Diffraction Imaging

AU - Yildiz, Orcun

AU - Chan, Henry

AU - Raghavan, Krishnan

AU - Judge, William

AU - Cherukara, Mathew J.

AU - Balaprakash, Prasanna

AU - Sankaranarayanan, Subramanian

AU - Peterka, Tom

PY - 2022

Y1 - 2022

N2 - X-ray Bragg coherent diffraction imaging (BCDI) is widely used for materials characterization. However, obtaining X-ray diffraction data is difficult and computationally intensive. Here, we introduce a machine learning approach to identify crystalline line defects in samples from the raw coherent diffraction data. To automate this process, we compose a workflow coupling coherent diffraction data generation with training and inference of deep neural network defect classifiers. In particular, we adopt a continual learning approach, where we generate training and inference data as needed based on the accuracy of the defect classifier instead of all training data generated a priori. The results show that our approach improves the accuracy of defect classifiers while using much fewer samples of data.

AB - X-ray Bragg coherent diffraction imaging (BCDI) is widely used for materials characterization. However, obtaining X-ray diffraction data is difficult and computationally intensive. Here, we introduce a machine learning approach to identify crystalline line defects in samples from the raw coherent diffraction data. To automate this process, we compose a workflow coupling coherent diffraction data generation with training and inference of deep neural network defect classifiers. In particular, we adopt a continual learning approach, where we generate training and inference data as needed based on the accuracy of the defect classifier instead of all training data generated a priori. The results show that our approach improves the accuracy of defect classifiers while using much fewer samples of data.

KW - HPC workflows

KW - catastrophic forgetting

KW - continual learning

KW - defect identification

UR - http://www.scopus.com/inward/record.url?scp=85147919493&partnerID=8YFLogxK

U2 - 10.1109/AI4S56813.2022.00006

DO - 10.1109/AI4S56813.2022.00006

M3 - Conference contribution

AN - SCOPUS:85147919493

T3 - Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis

SP - 1

EP - 6

BT - Proceedings of AI4S 2022

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 3rd IEEE/ACM International Workshop on Artificial Intelligence and Machine Learning for Scientific Applications, AI4S 2022

Y2 - 13 November 2022 through 18 November 2022

ER -

Yildiz O, Chan H, Raghavan K, Judge W, Cherukara MJ, Balaprakash P et al. Automated Continual Learning of Defect Identification in Coherent Diffraction Imaging. In Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis. Institute of Electrical and Electronics Engineers Inc. 2022. p. 1-6. (Proceedings of AI4S 2022: Artificial Intelligence and Machine Learning for Scientific Applications, Held in conjunction with SC 2022: The International Conference for High Performance Computing, Networking, Storage and Analysis). doi: 10.1109/AI4S56813.2022.00006

Automated Continual Learning of Defect Identification in Coherent Diffraction Imaging

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this