20 January 2022 in Journal of Electrical Engineering & Technology
Dissolved gas analysis (DGA) method is widely used to detect the incipient fault of power transformers. This paper presents a novel DGA method for power transformer fault diagnosis based on Harris-Hawks-optimization (HHO) algorithm optimized kernel extreme learning machine (KELM). The non-code ratios of the gases are used as the characterizing vector for the KELM model, and the Harris-Hawks-optimization (HHO) algorithm is introduced to optimize the KELM parameters, which promotes the fault diagnostic performance of KELM. Based on dataset collected from IEC TC 10, the fault diagnosis capability of the proposed method is validated by different characterizing vectors and is compared with conventional KELM and other optimized KELM. Moreover, the generalization ability of the proposed method is confirmed by China DGA data. The results demonstrate that the proposed method is superior to other methods and is more effective and stable for power transformer fault diagnosis with high accuracy.
6 June 2023 in Engineering Applications of Artificial Intelligence
9 June 2025 in ACM Transactions on Information Systems
Non-overlapping Cross-domain Sequential Recommendation (NCSR) is the task that focuses on domain knowledge transfer without overlapping entities. Compared with traditional Cross-domain Sequential Recommendation (CSR), NCSR poses several challenges: 1) NCSR methods often rely on explicit item IDs, overlooking semantic information among entities. 2) Existing CSR mainly relies on domain alignment for knowledge transfer, risking semantic loss during alignment. 3) Most previous studies do not consider the many-to-one characteristic, which is challenging because of the utilization of multiple source domains. Given the above challenges, we introduce the prompt learning technique for Many-to-one Non-overlapping Cross-domain Sequential Recommendation (MNCSR) and propose a Text-enhanced Co-attention Prompt Learning Paradigm (TCPLP). Specifically, we capture semantic meanings by representing items through text rather than IDs, leveraging natural language universality to facilitate cross-domain knowledge transfer. Unlike prior works that need to conduct domain alignment, we directly learn transferable domain information, where two types of prompts, i.e., domain-shared and domain-specific prompts, are devised, with a co-attention-based network for prompt encoding. Then, we develop a two-stage learning strategy, i.e., pre-train & prompt-tuning paradigm, for domain knowledge pre-learning and transferring, respectively. We conduct extensive experiments on three datasets and the experimental results demonstrate the superiority of our TCPLP. Our source codes have been publicly released 1 .
16 September 2025 in Measurement