ON THIS PAGE

Research on Risk Identification of Digital Voice Information in Electric Power Customer Service Center

Tongxu Liu1, Chuangye Wang1, Yu Wang2, Shilin Shen1, Yunchuan Wang1
1Bengbu Power Supply Company, State Grid Anhui Electric Power Co., Ltd. Bengbu 233000, Anhui, China
2Electric Power Research Institute, State Grid Anhui Electric Power Co., Ltd., Hefei 230601, Anhui, China

Abstract

With the increasing adoption of intelligent voice services in electric power customer service centers, identifying potential information risks from large-scale speech interactions has become essential for service compliance and information security management. However, existing speech-based risk identification methods often struggle with implicit semantic relationships, multi-turn contextual dependencies, and recognition errors under noisy acoustic conditions. To address these challenges, this paper proposes a digital speech information risk identification framework integrating Automatic Speech Recognition (ASR), Abstract Meaning Representation (AMR), and multi-granularity graph neural networks. Specifically, a dual-encoder architecture combines hierarchical Transformer-based textual representations with graph neural network-based AMR encoding to extract complementary contextual and structured semantic features. Clause-level and token-level adjacency structures, together with hierarchical self-attention and cross-attention fusion, are employed to capture long-range semantic dependencies and risk-related entity relationships. Furthermore, a Tacotron2–MelGAN-based speech augmentation module is incorporated to support diverse risk-oriented speech scenarios and robustness evaluation. The framework is designed to identify potential sensitive information disclosure, abnormal service statements, procedural violations, and contextual risks in customer service conversations. Comparative experiments involving SRL-BERT, DEP-GNN, and Lin-AMR, together with ablation studies, are designed to evaluate identification performance using Precision, Recall, Macro-F1, and AUPRC under realistic, simulated, and noise-augmented speech conditions.

Related Articles
Seda Evyapan Aydın1, Gökhan Aydın2
1Faculty of Sport Sciences, Uşak University, Uşak, Türkiye
2Department of Family Medicine, Afyonkarahisar Health Sciences University, Afyonkarahisar, Türkiye
Xide Wang1, Qingyan Fang2
1School of Artificial Intelligence, Harbin University, Harbin 150086, Heilongjiang, China
2School of Electrical and Intelligent Engineering, Harbin Guangsha College, Harbin 150001, Heilongjiang, China
Rahul Shah1
1The ICFAI University, Sikkim, Ranka Road, Lower Sichey, Gangtok, Sikkim 737101, India
Mustafa Salah Hasan1
1College of Veterinary Medicine, University of Fallujah, Fallujah 31002, Anbar, Iraq
Chuangye Wang1, Tongxu Liu1, Yu Wang2, Degui Qiu3, Shilin Shen1
1Bengbu Power Supply Company, State Grid Anhui Electric Power Co., Ltd., Bengbu 233000, Anhui, China
2Electric Power Research Institute, State Grid Anhui Electric Power Co., Ltd., Hefei 230601, Anhui, China
3State Grid Fengyang County Electric Power Supply Company, Fengyang 233100, Anhui, China

Citation

Tongxu Liu, Chuangye Wang, Yu Wang, Shilin Shen, Yunchuan Wang. Research on Risk Identification of Digital Voice Information in Electric Power Customer Service Center[J], Archives Des Sciences, Volume 76, Issue 5, 2026. 16-28. DOI: https://doi.org/10.68304/as/76502.