Enhancing Stress-Strain Predictions with Seq2Seq and Cross-Attention based on Small Punch Test

Zhengni Yang; Rui Yang; Weijian Han; Qixin Liu

doi:10.1109/IJCNN64981.2025.11228148

Enhancing Stress-Strain Predictions with Seq2Seq and Cross-Attention based on Small Punch Test

Zhengni Yang
, Rui Yang^*
, Weijian Han^*
, Qixin Liu

^*Corresponding author for this work

Xi'an Jiaotong-Liverpool University
Materials Bigdata and Applications Division Materials Academy JITRI
University of Liverpool

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

Abstract

This paper introduces a novel deep-learning approach to predict true stress-strain curves of high-strength steels from small punch test (SPT) load-displacement data. The proposed approach uses Gramian Angular Field (GAF) to transform load-displacement sequences into images, capturing spatial-temporal features and employs a Sequence-to-Sequence (Seq2Seq) model with an LSTM-based encoder-decoder architecture, enhanced by multi-head cross-attention to improved accuracy. Experimental results demonstrate that the proposed approach achieves superior prediction accuracy, with minimum and maximum mean absolute errors of 0.15 MPa and 5.58 MPa, respectively. The proposed method offers a promising alternative to traditional experimental techniques in materials science, enhancing the accuracy and efficiency of true stress-strain relationship predictions.

Original language	English
Title of host publication	International Joint Conference on Neural Networks, IJCNN 2025 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9798331510428
DOIs	https://doi.org/10.1109/IJCNN64981.2025.11228148
Publication status	Published - 2025
Event	2025 International Joint Conference on Neural Networks, IJCNN 2025 - Rome, Italy Duration: 30 Jun 2025 → 5 Jul 2025

Publication series

Name	Proceedings of the International Joint Conference on Neural Networks
ISSN (Print)	2161-4393
ISSN (Electronic)	2161-4407

Conference

Conference	2025 International Joint Conference on Neural Networks, IJCNN 2025
Country/Territory	Italy
City	Rome
Period	30/06/25 → 5/07/25

Keywords

Deep Learning
High-Strength Steels
Sequence-to-Sequence Model
Small Punch Test
Stress-Strain Curves

Access to Document

10.1109/IJCNN64981.2025.11228148

Cite this

Yang, Z., Yang, R., Han, W., & Liu, Q. (2025). Enhancing Stress-Strain Predictions with Seq2Seq and Cross-Attention based on Small Punch Test. In International Joint Conference on Neural Networks, IJCNN 2025 - Proceedings (Proceedings of the International Joint Conference on Neural Networks). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IJCNN64981.2025.11228148

@inproceedings{f7f50d6c0643444498bcde060fdf756b,

title = "Enhancing Stress-Strain Predictions with Seq2Seq and Cross-Attention based on Small Punch Test",

abstract = "This paper introduces a novel deep-learning approach to predict true stress-strain curves of high-strength steels from small punch test (SPT) load-displacement data. The proposed approach uses Gramian Angular Field (GAF) to transform load-displacement sequences into images, capturing spatial-temporal features and employs a Sequence-to-Sequence (Seq2Seq) model with an LSTM-based encoder-decoder architecture, enhanced by multi-head cross-attention to improved accuracy. Experimental results demonstrate that the proposed approach achieves superior prediction accuracy, with minimum and maximum mean absolute errors of 0.15 MPa and 5.58 MPa, respectively. The proposed method offers a promising alternative to traditional experimental techniques in materials science, enhancing the accuracy and efficiency of true stress-strain relationship predictions.",

keywords = "Deep Learning, High-Strength Steels, Sequence-to-Sequence Model, Small Punch Test, Stress-Strain Curves",

author = "Zhengni Yang and Rui Yang and Weijian Han and Qixin Liu",

note = "Publisher Copyright: {\textcopyright} 2025 IEEE.; 2025 International Joint Conference on Neural Networks, IJCNN 2025 ; Conference date: 30-06-2025 Through 05-07-2025",

year = "2025",

doi = "10.1109/IJCNN64981.2025.11228148",

language = "English",

series = "Proceedings of the International Joint Conference on Neural Networks",

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

booktitle = "International Joint Conference on Neural Networks, IJCNN 2025 - Proceedings",

}

Yang, Z, Yang, R, Han, W & Liu, Q 2025, Enhancing Stress-Strain Predictions with Seq2Seq and Cross-Attention based on Small Punch Test. in International Joint Conference on Neural Networks, IJCNN 2025 - Proceedings. Proceedings of the International Joint Conference on Neural Networks, Institute of Electrical and Electronics Engineers Inc., 2025 International Joint Conference on Neural Networks, IJCNN 2025, Rome, Italy, 30/06/25. https://doi.org/10.1109/IJCNN64981.2025.11228148

Enhancing Stress-Strain Predictions with Seq2Seq and Cross-Attention based on Small Punch Test. / Yang, Zhengni; Yang, Rui; Han, Weijian et al.
International Joint Conference on Neural Networks, IJCNN 2025 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2025. (Proceedings of the International Joint Conference on Neural Networks).

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

TY - GEN

T1 - Enhancing Stress-Strain Predictions with Seq2Seq and Cross-Attention based on Small Punch Test

AU - Yang, Zhengni

AU - Yang, Rui

AU - Han, Weijian

AU - Liu, Qixin

PY - 2025

Y1 - 2025

N2 - This paper introduces a novel deep-learning approach to predict true stress-strain curves of high-strength steels from small punch test (SPT) load-displacement data. The proposed approach uses Gramian Angular Field (GAF) to transform load-displacement sequences into images, capturing spatial-temporal features and employs a Sequence-to-Sequence (Seq2Seq) model with an LSTM-based encoder-decoder architecture, enhanced by multi-head cross-attention to improved accuracy. Experimental results demonstrate that the proposed approach achieves superior prediction accuracy, with minimum and maximum mean absolute errors of 0.15 MPa and 5.58 MPa, respectively. The proposed method offers a promising alternative to traditional experimental techniques in materials science, enhancing the accuracy and efficiency of true stress-strain relationship predictions.

AB - This paper introduces a novel deep-learning approach to predict true stress-strain curves of high-strength steels from small punch test (SPT) load-displacement data. The proposed approach uses Gramian Angular Field (GAF) to transform load-displacement sequences into images, capturing spatial-temporal features and employs a Sequence-to-Sequence (Seq2Seq) model with an LSTM-based encoder-decoder architecture, enhanced by multi-head cross-attention to improved accuracy. Experimental results demonstrate that the proposed approach achieves superior prediction accuracy, with minimum and maximum mean absolute errors of 0.15 MPa and 5.58 MPa, respectively. The proposed method offers a promising alternative to traditional experimental techniques in materials science, enhancing the accuracy and efficiency of true stress-strain relationship predictions.

KW - Deep Learning

KW - High-Strength Steels

KW - Sequence-to-Sequence Model

KW - Small Punch Test

KW - Stress-Strain Curves

UR - https://www.scopus.com/pages/publications/105023968182

U2 - 10.1109/IJCNN64981.2025.11228148

DO - 10.1109/IJCNN64981.2025.11228148

M3 - Conference Proceeding

AN - SCOPUS:105023968182

T3 - Proceedings of the International Joint Conference on Neural Networks

BT - International Joint Conference on Neural Networks, IJCNN 2025 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2025 International Joint Conference on Neural Networks, IJCNN 2025

Y2 - 30 June 2025 through 5 July 2025

ER -

Yang Z, Yang R, Han W, Liu Q. Enhancing Stress-Strain Predictions with Seq2Seq and Cross-Attention based on Small Punch Test. In International Joint Conference on Neural Networks, IJCNN 2025 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2025. (Proceedings of the International Joint Conference on Neural Networks). doi: 10.1109/IJCNN64981.2025.11228148

Enhancing Stress-Strain Predictions with Seq2Seq and Cross-Attention based on Small Punch Test

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Cite this