Qiming Li, Xiaocheng Feng, Yixuan Ma, Ruihan Chen, Zihe Tong, Zekai Ye, Xiachong Feng, Libo Qin, Haoyu Ren, Kun Chen, Yunfei Lu, Dandan Tu, Bing Qin. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Hypergraphs effectively model higher-order relationships in natural phenomena, capturing complex interactions beyond pairwise connections. We introduce a novel hypergraph message passing framework inspired by interacting particle systems, where hyperedges act as fields inducing shared node dynamics. By incorporating attraction, repulsion, and Allen-Cahn forcing terms, particles of varying classes and features achieve class-dependent equilibrium, enabling separability through the particle-driven message passing. We investigate both first-order and second-order particle system equations for modeling these dynamics, which mitigate over-smoothing and heterophily thus can capture complete interactions. The more stable second-order system permits deeper message passing. Furthermore, we enhance deterministic message passing with stochastic element to account for interaction uncertainties. We prove theoretically that our approach mitigates over-smoothing by maintaining a positive lower bound on the hypergraph Dirichlet energy during propagation and thus to enable hypergraph message passing to go deep. Empirically, our models demonstrate competitive performance on diverse real-world hypergraph node classification tasks, excelling on both homophilic and heterophilic datasets. Source code is available at \href{https://github.com/Xuan-Elfin/HAMP}{the link}.

Accurate pantograph-catenary attitude detection is essential for high-speed railway safety, yet existing methods face significant challenges when dealing with sparse keypoints, small target regions, and complex operational environments. To address these limitations, we propose PantoPoseNet, a novel two-stage framework designed for real-time pantograph-catenary keypoint detection. Our approach introduces three key innovations in the first stage: (1) integration of VanillaNet with SIAF activation as the backbone network, achieving a 55.6% reduction in model parameters while preserving detection accuracy; (2) replacement of the conventional RepC3 module with CSP modules to enhance multiscale feature fusion capabilities; and (3) implementation of a hybrid loss function that combines GIoU and NWD metrics, specifically engineered to mitigate gradient vanishing issues inherent in small keypoint region detection. The second stage employs a specialized VanillaNet-KP network that processes 32×32 pixel regions to achieve precise keypoint localization. Comprehensive experiments conducted across various railway operational scenarios demonstrate that PantoPoseNet achieves superior performance with 98.2% mAP for keypoint region detection and 93.6% PCK for overall keypoint localization, while maintaining real-time processing at 25.94 FPS. These results significantly outperform current state-of-the-art methods, indicating strong potential for practical deployment in pantograph-catenary monitoring systems.

Zhouhao Sun, Li Du, Xiao Ding, Yixuan Ma, Yang Zhao, Kaitao Qiu, Ting Liu, Bing Qin. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

Reliable and efficient arc detection is critical for ensuring the safe operation of high-speed railways. However, existing methods often suffer from low accuracy and high computational complexity, hindering their practical application. To tackle these challenges, we propose a novel deep learning-based method that significantly improves the YOLOv5 object detection network for real-time arc detection in pantograph-catenary systems. Our approach introduces two key innovations: (1) a Squeeze-and-Excitation-based C3 (SEC3) attention module to adaptively prioritize informative features, and (2) a Bi-directional Feature Pyramid Network (BiFPN) for enhanced multi-scale feature fusion. We conduct comprehensive experiments on a diverse dataset collected from real-world scenarios, covering various challenging conditions. The results show that our improved YOLOv5 network achieves superior performance compared to state-of-the-art methods in both accuracy and efficiency. Moreover, ablation studies confirm the merits of the proposed SEC3 attention and BiFPN fusion modules in boosting performance. Our approach offers a promising solution for automatic arc detection, thereby contributing to the enhanced safety and reliability of high-speed railway systems.

With the rapid expansion of freelance workers and tasks, online labor market platforms face a significant challenge with the cold-start problem, which makes it very difficult to effectively match new workers with suitable tasks. To solve this challenge, this paper presents a novel solution termed the Meta-learning ID Embedding Generator (MIG). MIG addresses the cold-start problem in task recommendation systems by efficiently learning suitable ID embeddings for new workers. MIG consists of an initial embedding generator for generating the initial ID embedding, alongside two adaptors designed to iteratively refine this embedding on the worker's competence and interest. The efficacy of this method has been assessed using authentic data sourced from Freelancer.com, a leading online labor marketplace. The empirical findings demonstrate its superiority over state-of-the-art methods when addressing the needs of two challenging user segments: newcomers to the platform and long-inactive users whose bidding records are sparse. MIG can seamlessly integrate into existing task recommendation systems, thereby enhancing their effectiveness, particularly in cold start scenarios.

Augmentation techniques and sampling strategies are crucial in contrastive learning, but in most existing works, augmentation techniques require careful design, and their sampling strategies can only capture a small amount of intrinsic supervision information. Additionally, the existing methods require complex designs to obtain two different representations of the data. To overcome these limitations, we propose a novel framework called the Self-Contrastive Graph Diffusion Network (SCGDN). Our framework consists of two main components: the Attentional Module (AttM) and the Diffusion Module (DiFM). AttM aggregates higher-order structure and feature information to get an excellent embedding, while DiFM balances the state of each node in the graph through Laplacian diffusion learning and allows the cooperative evolution of adjacency and feature information in the graph. Unlike existing methodologies, SCGDN is an augmentation-free approach that avoids "sampling bias" and semantic drift, without the need for pre-training. We conduct a high-quality sampling of samples based on structure and feature information. If two nodes are neighbors, they are considered positive samples of each other. If two disconnected nodes are also unrelated on kNN graph, they are considered negative samples for each other. The contrastive objective reasonably uses our proposed sampling strategies, and the redundancy reduction term minimizes redundant information in the embedding and can well retain more discriminative information. In this novel framework, the graph self-contrastive learning paradigm gives expression to a powerful force. The results manifest that SCGDN can consistently generate out performance over both the contrastive methods and the classical methods. The source code is available at https://github.com/kunzhan/SCGDN.

Contrastive learning on graphs aims at extracting distinguishable high-level representations of nodes. We theoretically illustrate that the entropy of a dataset is approximated by maximizing the lower bound of the mutual information across different views of a graph, i.e., entropy is estimated by a neural network. Based on this finding, we propose a simple yet effective subset sampling strategy to contrast pairwise representations between views of a dataset. In particular, we randomly sample nodes and edges from a given graph to build the input subset for a view. Two views are fed into a parameter-shared Siamese network to extract the high-dimensional embeddings and estimate the information entropy of the entire graph. For the learning process, we propose to optimize the network using two objectives, simultaneously. Concretely, the input of the contrastive loss consists of positive and negative pairs. Our selection strategy of pairs is different from previous works and we present a novel strategy to enhance the representation ability by selecting nodes based on cross-view similarities. We enrich the diversity of the positive and negative pairs by selecting highly similar samples and totally different data with the guidance of cross-view similarity scores, respectively. We also introduce a cross-view consistency constraint on the representations generated from the different views. We conduct experiments on seven graph benchmarks, and the proposed approach achieves competitive performance compared to the current state-of-the-art methods. The source code is available at https://github.com/kunzhan/M-ILBO.

Although many graph-based clustering methods attempt to model the stationary diffusion state in their objectives, their performance limits to using a predefined graph. We argue that the estimation of the stationary diffusion state can be achieved by gradient descent over neural networks. We specifically design the Stationary Diffusion State Neural Estimation (SDSNE) to exploit multiview structural graph information for co-supervised learning. We explore how to design a graph neural network specially for unsupervised multiview learning and integrate multiple graphs into a unified consensus graph by a shared self-attentional module. The view-shared self-attentional module utilizes the graph structure to learn a view-consistent global graph. Meanwhile, instead of using auto-encoder in most unsupervised learning graph neural networks, SDSNE uses a co-supervised strategy with structure information to supervise the model learning. The co-supervised strategy as the loss function guides SDSNE in achieving the stationary state. With the help of the loss and the self-attentional module, we learn to obtain a graph in which nodes in each connected component fully connect by the same weight. Experiments on several multiview datasets demonstrate effectiveness of SDSNE in terms of six clustering evaluation metrics.

Driven by the high profit, Portable Executable (PE) malware has been consistently evolving in terms of both volume and sophistication. PE malware family classification has gained great attention and a large number of approaches have been proposed. With the rapid development of machine learning techniques and the exciting results they achieved on various tasks, machine learning algorithms have also gained popularity in the PE malware family classification task. Three mainstream approaches that use learning based algorithms, as categorized by the input format the methods take, are image-based, binary-based and disassembly-based approaches. Although a large number of approaches are published, there is no consistent comparisons on those approaches, especially from the practical industry adoption perspective. Moreover, there is no comparison in the scenario of concept drift, which is a fact for the malware classification task due to the fast evolving nature of malware. In this work, we conduct a thorough empirical study on learning-based PE malware classification approaches on 4 different datasets and consistent experiment settings. Based on the experiment results and an interview with our industry partners, we find that (1) there is no individual class of methods that significantly outperforms the others; (2) All classes of methods show performance degradation on concept drift (by an average F1-score of 32.23%); and (3) the prediction time and high memory consumption hinder existing approaches from being adopted for industry usage.