Improving large language models (LLMs) for electronic health record (EHR) reasoning is essential for enabling accurate and generalizable clinical predictions. While LLMs excel at medical text understanding, they underperform on EHR-based prediction tasks due to challenges in modeling temporally structured, high-dimensional data. Existing approaches often rely on hybrid paradigms, where LLMs serve merely as frozen prior retrievers while downstream deep learning (DL) models handle prediction, failing to improve the LLM’s intrinsic reasoning capacity and inheriting the generalization limitations of DL models. To this end, we propose EAG-RL, a novel two-stage training framework designed to intrinsically enhance LLMs’ EHR reasoning ability through expert attention guidance, where expert EHR models refer to task-specific DL models trained on EHR data. Concretely, EAG-RL first constructs high-quality, stepwise reasoning trajectories using expert-guided Monte Carlo Tree Search to effectively initialize the LLM’s policy. Then, EAG-RL further optimizes the policy via reinforcement learning by aligning the LLM’s attention with clinically salient features identified by expert EHR models. Extensive experiments on two real-world EHR datasets show that EAG-RL improves the intrinsic EHR reasoning ability of LLMs by an average of 14.62%, while also enhancing robustness to feature perturbations and generalization to unseen clinical domains. These results demonstrate the practical potential of EAG-RL for real-world deployment in clinical prediction tasks.

Zhibang Yang, Xinke Jiang, Rihong Qiu, Ruiqing Li, Yihang Zhang, Yue Fang, Yongxin Xu, Hongxin Ding, Xu Chu, Junfeng Zhao, Yasha Wang. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Graph pattern queries (GPQ) over RDF graphs extend basic graph patterns to support variable-length paths (VLP), thereby enabling complex knowledge retrieval and navigation. Generally, variable-length paths describe the reachability between two vertices via a given property within a specified range. With the increasing scale of RDF graphs, it is necessary to design a partitioning method to achieve efficient distributed queries. Although many partitioning strategies have been proposed for large RDF graphs, most existing methods result in numerous inter-partition joins when processing GPQs, which impacts query performance. In this paper, we formulate a new partitioning problem, MaxLocJoin, aims to minimize inter-partition joins during distributed GPQ processing. For MaxLocJoin, we propose a partitioning framework (PIP) based on property-induced subgraphs, which consist of edges with a specific set of properties. The framework first finds a locally joinable property set using a cost-driven algorithm, LJPS, where the cost depends on the sizes of weakly connected components within its property-induced subgraphs. Subsequently, the graph is partitioned according to the weakly connected components. The framework can achieve two key objectives: first, it enables complete local processing of all variable-length path queries (eliminating inter-partition joins); second, it can minimize the number of inter-partition joins required for traditional graph pattern queries. Moreover, we identify two types of independently executable queries (IEQ): the locally joinable IEQ and the single-property IEQ. After that, a query decomposition algorithm is designed to transform all GPQ into one of them for independent execution in distributed environments. In experiments, we implement two prototype systems based on Jena and Virtuoso, and evaluate them over both real and synthetic RDF graphs. The results show that MaxLocJoin achieves performance improvements from 2.8x to 10.7x over existing methods.

Spatial crowdsourcing (SC) is becoming increasingly popular recently. As a critical issue in SC, task assignment currently faces challenges due to the imbalanced spatiotemporal distribution of tasks. Hence, many related studies and applications focusing on cross-platform task allocation in SC have emerged. Existing work primarily focuses on the maximization of total revenue for inner platform in cross task assignment. In this work, we formulate a SC problem called Cross Dynamic Task Assignment (CDTA) to maximize the overall utility and propose improved solutions aiming at creating a win-win situation for inner platform, task requesters, and outer workers. We first design a hybrid batch processing framework and a novel cross-platform incentive mechanism. Then, with the purpose of allocating tasks to both inner and outer workers, we present a KM-based algorithm that gets the accurate assignment result in each batch and a density-aware greedy algorithm with high efficiency. To maximize the revenue of inner platform and outer workers simultaneously, we model the competition among outer workers as a potential game that is shown to have at least one pure Nash equilibrium and develop a game-theoretic method. Additionally, a simulated annealing-based improved algorithm is proposed to avoid falling into local optima. Last but not least, since random thresholds lead to unstable results when picking tasks that are preferentially assigned to inner workers, we devise an adaptive threshold selection algorithm based on multi-armed bandit to further improve the overall utility. Extensive experiments demonstrate the effectiveness and efficiency of our proposed algorithms on both real and synthetic datasets.

As the network data in real life become multi-modal and multi-relational, multi-view attributed graphs have garnered significant attention. Numerous methods have achieved excellent performance in multi-view attributed graph clustering; however, they cannot efficiently handle incomplete attribute scenarios, which are prevalent in many real-life applications. Inspired by this, we investigate the problem of multi-view incomplete attributed graph clustering for the first time. In particular, the TOTF (Train Once Then Freeze) framework is designed to train missing-aware encoders that capture view-specific information while ignoring the impact of incomplete attributes, and then employs frozen encoders to uncover common information driven by clustering. After that, we propose a correlation strength-aware graph neural network on the basis of the inherent relationships among attributes to enhance accuracy. It is proven theoretically that traditional Generative Adversarial Networks (GANs) are unable to generate the unique real distribution. To address this issue, we further introduce the missing-position reminder mechanism into our intra-view adversarial games for better clustering results. Extensive experimental results demonstrate that our method achieves up to a 17% improvement in accuracy over the state-of-the-art methods. The source code is available at https://anonymous.4open.science/r/TOTF-main.

Truth discovery endeavors to extract valuable information from multi-source data through weighted aggregation. Some studies have integrated differential privacy techniques into traditional truth discovery algorithms to protect data privacy. However, due to the neglect of outliers and limitations in budget allocation, these schemes still need improvement in the accuracy of discovery results. To solve these challenges, we propose a privacy-preserving scheme called PriPTD to achieve secure and accurate truth discovery services over crowdsourced data streams. Instead of assuming that worker weights are always stable between two neighboring timestamps, we delve deeper to consider outliers where worker weights change rapidly. Accordingly, we develop an outlier-aware weight estimation method with a time series model to capture and handle these outliers. Furthermore, to ensure data utility under a limited budget, we devise a weight-aware budget allocation algorithm. Its core idea is that timestamps with higher importance consume a larger proportion of the remaining budget. Additionally, we design a noise-aware error adjustment approach to mitigate the adverse effects of introduced noise on accuracy. Theoretical analysis and extensive experiments validate our scheme. Final comparative experiments against existing works confirm that our scheme achieves more accurate truth discovery while preserving privacy.

Due to the limited computing resources of both mobile devices (MDs) and the mobile edge computing (MEC) server, devising reasonable strategies for MD task offloading, MEC server resource pricing, and resource allocation is crucial. In this paper, a scenario is considered, comprising multiple MDs and a single MEC server. Each MD has a divisible task in each time slot, allowing for partial offloading and the option to discard parts of the task. The MEC server contains multiple computing units with the same computing power, and its computing resources can be dynamically adjusted through dynamic voltage and frequency scaling (DVFS) according to the size of tasks offloaded by MDs. At any given time slice, a Stackelberg game is formulated based on the strategies of the MDs and the strategy of the MEC server. An iterative evolution algorithm is employed to explore the optimal strategies for MDs and the MEC server. Simulation results demonstrate that both parties can reach an equilibrium state through the game, and these experiments confirm that the algorithm effectively enhances system efficiency.

Continuous subgraph matching (CSM) is an important problem of graph analysis over dynamic graphs. Given a query graph, existing CSM efforts return numerous matches, which may overwhelm analysts. In addition, they do not consider weighted graphs that are ubiquitous in many real-world applications, such as payment networks where each edge has a weight to represent a transaction amount. Over these weighted graphs, matches of the given query graph have diverse priorities for analysis if they own different densities. In this paper, we propose a new problem of CSM-TopK to compute$k$matches of a given query graph with the highest densities over a dynamic weighted graph and prove it to be NP-hard. To compute the CSM-TopK effectively, we first define a star-structured subquery, based on which we design two lightweight indexes, called global and local MWstar, respectively. In particular, the global MWstar maintains the maximum weights of all partial matches of each specific star-structured subquery. Differently, the local MW star is designed based on the corresponding maximum weight distribution for each specific data vertex. Additionally, a query-dependent graph compacted technique is introduced to further improve the performance on both time and space. Extensive experiments over real-world datasets show that our MW star- based approaches surpass the comparative ones by at least two orders of magnitude.

The Multimodal Sentiment Classification (MSC) task aims to discern sentiments from diverse data sources. Existing efforts focus on integrating multimodal features and enhancing representation learning for improved recognition. The widespread use of MSC, particularly in risk-associated domains, highlights the need for heightened trustworthiness in predictions. However, most current MSC models often provide elevated confidence regardless of whether the prediction is correct or not, with less emphasis on whether this confidence reasonably reflects the model’s certainty in predictions. This paper proposes a novel confidence optimization framework, Sentiment Confidence Separation (SCS), which helps address unreliability in MSC models by making the correct and incorrect predictions output discriminative confidences. SCS comprises Confidence Separation Loss (CSL) and Flatness-Based Separation Optimization (FBSO), facilitating reliable and precise predictions. Comprehensive experimentation validates the efficacy of the proposed approach across multiple mainstream datasets.

With the increasingly powerful performances and enormous scales of pretrained models, promoting parameter efficiency in fine-tuning has become a crucial need for effective and efficient adaptation to various downstream tasks. One representative line of fine-tuning methods is Orthogonal Fine-tuning (OFT), which rigorously preserves the angular distances within the parameter space to preserve the pretrained knowledge. Despite the empirical effectiveness, OFT still suffers low parameter efficiency at $\mathcal{O}(d^2)$ and limited capability of downstream adaptation. Inspired by Givens rotation, in this paper, we proposed quasi-Givens Orthogonal Fine-Tuning (qGOFT) to address the problems. We first use $\mathcal{O}(d)$ Givens rotations to accomplish arbitrary orthogonal transformation in $SO(d)$ with provable equivalence, reducing parameter complexity from $\mathcal{O}(d^2)$ to $\mathcal{O}(d)$. Then we introduce flexible norm and relative angular adjustments under soft orthogonality regularization to enhance the adaptation capability of downstream semantic deviations. Extensive experiments on various tasks and pretrained models validate the effectiveness of our methods.

One area of current research on adversarial attacks is how to generate plausible adversarial examples when only a small number of datasets are available. Current adversarial attack algorithms used to attack these black-box systems face a number of challenges, such as difficulty in training convergence, ambiguous sample images, substitute models collapse, unsatisfactory attack success rates, high query cost, and low defense capability improvement of target models. As a result, constructing plausible adversarial situations in a few known real-world sample circumstances remains difficult. As a solution to the aforementioned issues, this study introduces MC-Net, a novel multi-stage and multi-class balanced generating method based on a limited number of samples to generate realistic adversarial examples. Firstly, a multi-task learning approach is used to train the GAN by fully utilizing the small samples, ensuring that the size of the generated dataset for each category is balanced. In addition, we design a weight-balancing strategy to ensure faster convergence of each sub-network. Then, in the second stage, the generated samples of different categories are used to train a substitute model, and the distillation method is adopted to learn the output distribution of the target model. Finally, adversarial examples are constructed on the generated samples to complete the attack on the target models. Extensive experiments have proven that MC-Net has the following advantages: 1) The substitute model converges quickly using limited samples and queries; 2) High attack success rates can be obtained with a few queries; and 3) The constructed adversarial examples significantly improve the target model’s defense. Furthermore, we only utilize a few queries for the Microsoft Azure online model to obtain a satisfactory result. Our code can be found at https://github.com/jiaokailun/A-fast.

Anomaly detection for multivariate time series is an essential task in the modern industrial field. Although several methods have been developed for anomaly detection, they usually fail to effectively exploit the metrical-temporal correlation and the other dependencies among multiple variables. To address this problem, we propose a stacked attention autoencoder for anomaly detection in multivariate time series (SA2E-AD); it focuses on fully utilizing the metrical and temporal relationships among multivariate time series. We design a multiattention block, alternately containing the temporal attention and metrical attention components in a hierarchical structure to better reconstruct normal time series, which is helpful in distinguishing the anomalies from the normal time series. Meanwhile, a two-stage training strategy is designed to further separate the anomalies from the normal data. Experiments on three publicly available datasets show that SA2E-AD outperforms the advanced baseline methods in detection performance and demonstrate the effectiveness of each part of the process in our method.

Cyber-physical systems (CPSs) are intricate systems integrating both physical and computational components. When these components fail due to malfunction or cyber-attack, potentially leading to significant damage or even collapse of the network topology. Cyber resilience, defined as the capability of a network to restore its function and structure after component failures, is crucial for ensuring that CPSs can sustain their operational capabilities in the face of complex disturbances. Recently, CPS resilience has garnered increasing attention, leading to the development of various resilience recovery methods. However, most existing studies address network and physical layer resilience in isolation, which hampers the ability to implement adaptive resilience recovery decisions across different systems. To overcome these limitations, we propose a multi-layered cascaded resilient recovery framework grounded in deep reinforcement learning. Initially, we synthesize the complex interactions between the information and physical layers in CPS resilience recovery from a global perspective, modeling the interrelations within CPSs. Subsequently, we introduce a hybrid resilient recovery strategy, encompassing both horizontal and vertical resilient recovery. The correlation matrix is used to partition the system into horizontal and vertical resilience slices. The resilient recovery strategy is subsequently modeled as an optimization problem using these slices. Following this, the Deep Recurrent Q-learning (DRQL) algorithm is introduced to implement the resilient recovery strategy in CPSs. While DRQL exhibits strong adaptability, it may lead to the sparse selection of critical samples, thereby hindering the learning process and convergence on essential experiences. To address this issue, we further develop the RR-DRQL algorithm, designed to identify the optimal CPS resilient recovery strategy. The RR-DRQL algorithm is rigorously proven to converge to the optimal solution through extensive theoretical analysis. Comprehensive experiments demonstrate that the RR-DRQL algorithm surpasses existing resilience recovery methods by 3.8%–25% regarding resilient policy recovery performance across realistic scenarios and various simulation platforms.

Summary With the rise of industrialization, the importance of the industrial Internet of Things (IIoT) has increased significantly, and with it comes a variety of security threats. Therefore, the security of these networks is critical. Industrial Response Systems (IRSs), as the last line of security, plays an important role in the security system of the Industrial Internet of Things. In this paper, a new IRS model based on the non‐cooperative game is proposed. First, by combining the Partially Observable Markov Decision Process (POMDP) model with the stochastic game model based on the expanded attack tree, our model could effectively perceive the changes at each node. Second, our model incorporates the alarms of intrusion detection system (IDS) and the physical quantities of sensors in Industrial Cyber‐Physical System (ICPS) into the quantization system so that the model can respond to intruders more accurately and comprehensively. Finally, we develop this model based on multiprocessors to speed up the solution process, and adopt an approximation algorithm to reduce the number of iterations of the POMDP

The ongoing digitalization enables stable control processes and smooth operations of Industrial Control Systems (ICSs). A direct consequence of the highly interconnected architecture of ICSs is the introduced cyber vulnerability and increasing cyber security threats to ICSs. Numerous researches pay attention to the security problem of ICSs. However, most current researches face two challenges. First, the interaction problem between cyber layer and physical layer of ICSs may result incorrect attack response strategies. Second, ICSs are real-time systems, but existing defense decision algorithms based on game theory or reinforcement learning techniques have high computational complexity, which prevents it from making decisions quickly. In this paper, we design a new multi-attribute based reward quantitative method and propose a multi-attribute based Q-learning algorithm to resolve the interaction problem. In addition, to overcome the limitation of slow convergence, we develop an effective parallel Q-learning (PQL) algorithm to quickly find the optimal strategy. The experimental results show the effectiveness of the PQL algorithm. Compared with the Q-learning algorithm (QL) and the deep Q-network (DQN) algorithm, our proposed solution can reduce the average completion time by 12.5%-37%.

Modular exponentiation, an operation widely utilized in cryptographic protocols to transfer text and other forms of data, can also be applied to Internet-of-Things (IoT) devices with high security requirements. However, due to the high resource consumption of modular exponentiation, IoT devices can face the problem of resource insufficient. Fortunately, the secure outsourcing scheme offers a new solution for resource-constrained devices. In this article, we apply a parallel secure outsourcing scheme to provide the possibility for modular exponentiation operation, which is used in the IoT devices. After that, the task of modular exponentiation is decomposed and we introduce the scheme in more detail. In addition, based on this scheme, we designed an extension scheme for RSA, providing enhanced security for IoT devices. Finally, the analysis of experimental results based on 512-4096 b of data indicates the superiority in scalability and time consumption over the previous schemes.

The G-Skyline (GSky) query is a powerful tool to analyze optimal groups in decision support. Compared with other group skyline queries, it releases users from providing an aggregate function. Besides, it can get much comprehensive results without overlooking some important results containing non-skylines. However, it is hard for the users to make sensible choices when facing so many results the GSky query returns, especially over a large, high-dimensional dataset or with a large group size. In this article, we investigate k representative G-Skyline ( k GSky) queries to obtain a manageable size of optimal groups. The k GSky query can also inherit the advantage of the GSky query; its results are representative and diversified. Next, we propose three exact algorithms with novel techniques including an upper bound pruning, a grouping strategy, a layered optimum strategy, and a hybrid strategy to efficiently process the k GSky query. Consider these exact algorithms have high time complexity and the precise results are not necessary in many applications. We further develop two approximate algorithms to trade off some accuracy for efficiency. Extensive experiments on both real and synthetic datasets demonstrate the efficiency, scalability, and accuracy of the proposed algorithms.

Nowadays, with the development of e-commerce, a growing number of customers choose to go shopping online. To find attractive products from online shopping marketplaces, the skyline query is a useful tool which offers more interesting and preferable choices for customers. The skyline query and its variants have been extensively investigated. However, to the best of our knowledge, they have not taken into account the requirements of customers in certain practical application scenarios. Recently, online shopping marketplaces usually hold some price promotion campaigns to attract customers and increase their purchase intention. Considering the requirements of customers in this practical application scenario, we are concerned about product selection under price promotion. We formulate a constrained optimal product combination (COPC) problem. It aims to find out the skyline product combinations which both meet a customer's willingness to pay and bring the maximum discount rate. The COPC problem is significant to offer powerful decision support for customers under price promotion, which is certified by a customer study. To process the COPC problem effectively, we first propose a two list exact (TLE) algorithm. The COPC problem is proven to be NP-hard, and the TLE algorithm is not scalable because it needs to process an exponential number of product combinations. Additionally, we design a lower bound approximate (LBA) algorithm that has a guarantee about the accuracy of the results and an incremental greedy (IG) algorithm that has good performance. The experiment results demonstrate the efficiency and effectiveness of our proposed algorithms.

Group skyline query is a powerful tool for optimal group analysis. Most of the existing group skyline queries select optimal groups by comparing the dominance relationship between aggregate-based points; such feature creates difficulties for users to specify an appropriate aggregate function. Besides, many significant groups that have great attractions to users in practice may be overlooked. To address these issues, the group skyline (GSky) query is formulated on the basis of a general definition of group dominance operator. While the existing GSky query algorithms are effective, there is still room for improvement in terms of progressiveness and efficiency. In this paper, we propose some new lemmas which facilitate direct generation of the GSky query results. Consecutively, we design a layered unit-based (LU) algorithm that applies a layered optimum strategy. Additionally, for the GSky query over the data that are dynamically produced and cannot be indexed, we propose a novel index-independent algorithm, called sorted-based progressive (SP) algorithm. The experimental results demonstrate the effectiveness, efficiency, and progressiveness of the proposed algorithms. By comparing with the state-of-the-art algorithm for the GSky query, our LU algorithm is more scalable and two orders of magnitude faster.

Nowadays, department stores and online merchants usually develop some price promotion strategies to attract customers and increase their purchase intention. Therefore, it is significant for customers to pick out attractive products and obtain the maximum discount rate. Admittedly, the skyline query is a most useful tool to find out attractive products. However, it does little to help select the product combinations with the maximum discount rate. Motivated by this, we identify an interesting problem, a most preferential skyline product (MPSP) combination discovering problem, which is NP-hard, for the first time in the literature. This problem aims to report all skyline product combinations having the maximum discount rate. Since the exact algorithm for the MPSP is not scalable to large or high-dimensional datasets, we design an approximate algorithm that guarantees the accuracy of the results. The experiment results demonstrate the efficiency and effectiveness of our proposed algorithms.