Volume 152
Published on June 2025Volume title: Proceedings of the 3rd International Conference on Software Engineering and Machine Learning
The rapid advancement of Internet-based healthcare technologies drives the daily generation of massive medical datasets, which hold substantial value for enhancing clinical decision support systems and facilitating evidence-based real-world medical research. Medical named entity recognition (NER) is important in the aforementioned research topics. In this paper, we propose a novel Bidirectional Gated Pyramid Network (BGPN), which consists of a convolution layer for extracting character-level features, a bidirectional LSTM (BiLSTM) layer for extracting local inter-sentence information, a Transformer layer for extracting long-distance textual information, and a gated fusion layer for dynamically updating the fusion weights of different levels. In addition, we incorporate a conditional random field (CRF), which enables the network to output the optimal prediction sequence of the BIO label. We validate our proposed method on the BC5CDR dataset, and the results show that our model achieves F1 scores of 0.79 and 0.70 for the two classes of named entities, chemical, and disease, with accuracies of 0.91 and 0.71, respectively.
View pdf
Decentralized Finance (DeFi) has revolutionized financial transactions by enabling open, permissionless access to financial services. However, its lack of centralized oversight and pseudonymous architecture have also brought by fraudulent activities. This study presents a novel framework for fraud detection in DeFi that integrates graph neural networks (GNNs) with multi-agent reinforcement learning (MARL). Leveraging a directed transaction graph comprising 50,000 Ethereum addresses and over 120,000 token transfers, this paper evaluates four detection pipelines: extreme gradient-boosted decision trees (XGBoost), a GNN-only model (GCN), a standalone reinforcement learning agent (PPO), and a proposed GNN+RL hybrid model. The hybrid system combines graph-based embeddings with adversarial policy learning, where a fraudster and a detector co-evolve through a multi-agent PPO setup using PettingZoo’s ParallelEnv. Synthetic fraud strategies are generated using a GAN and projected into the GCN embedding space to simulate adaptive threats. Experimental results show that while GCNs outperform flat-feature models, the GNN+RL hybrid achieves superior balance across accuracy (84.58%), AUC (0.8176), and F1 score (0.7493), capturing both structural and behavioral fraud signals. Reward convergence curves further illustrate emergent adversarial dynamics. The proposed framework demonstrates the effectiveness of combining relational inductive biases, dynamic decision-making, and adversarial augmentation for resilient fraud detection. Future work includes extending to cross-chain analytics and enriching contextual understanding through integration with large language models.
View pdf
Brain-Computer Interface (BCI) technology has become one of the focal points of scientific research in the modern world. Researchers from all corners of the globe have made significant contributions to this field, driving the continuous advancement of BCI technology. By 2025, Brain-Computer Interface (BCI) technology has become a groundbreaking human-machine interaction technology, enabling communication between the human brain and the external world, no longer a fantasy from science fiction but a tangible reality. It has now been widely applied across numerous fields, including communication, movement control, environment control, neurorehabilitation, and others.This paper first presents the basic principles of brain-computer interface (BCI) technology and provides a detailed classification of BCIs. It then reviews and enumerates the latest advancements in BCI research, exploring the various steps that form a standard BCI, including signal acquisition, preprocessing, feature extraction, and the control interface, while also investigating several mathematical algorithms. The aim is to gain a deeper understanding of BCI technology. Finally, the paper concludes with a summary and outlook on the future development prospects of this technology.
View pdf
In September 2020, the Chinese government pledged at the 75th United Nations General Assembly to achieve peak carbon emissions before 2030 and carbon neutrality before 2060. As the nation advances its carbon peaking and carbon neutrality goals, the demand for technologies supporting green and low-carbon transitions has surged, particularly in high-energy-consuming sectors like power inspection. Traditional manual inspection methods are struggling to meet the operational and maintenance demands of modern smart grids due to limitations such as low efficiency, high operational risks, and incomplete inspection coverage. While unmanned aerial vehicle (UAV) inspection offers significant advantages, its limited flight time remains a major challenge to its widespread adoption. This paper focuses on analyzing and optimizing the endurance of UAVs in power line inspection based on wireless power transfer (WPT) technology. The aim is to provide a more efficient charging solution for UAVs by proposing strategies to maximize channel gain, thereby overcoming endurance limitations and promoting the extensive application of UAV technology in smart grid inspection to support the national strategy for green and low-carbon transition.
View pdf
