Large Language Models (LLMs) have demonstrated remarkable In-Context learning (ICL) capabilities for relation extraction (RE). While ICL has shown promise in RE tasks, current approaches face challenges in example selection and utilization. These challenges stem from the misalignment between example selection methods and LLMs' inherent cognitive processing mechanisms, particularly in pattern recognition and relational reasoning. To address these limitations, we propose Counterfactual Cognitive Alignment (CCA), a novel framework that systematically enhances ICL performance in RE by aligning example selection with cognitive principles underlying human relational reasoning. The framework incorporates a cognitive-inspired counterfactual generation mechanism that creates semantically diverse yet relationally coherent examples, mirroring human "what-if" reasoning processes. Additionally, it employs a cognitive alignment approach that integrates structural identification features with semantic understanding to better align with LLMs cognitive processing patterns. Extensive experiments across multiple RE benchmarks reveal the effectiveness of our cognitive alignment approach through the synergistic integration of counterfactual reasoning and cognitively-guided selection.

As vehicular ad hoc networks (VANETs) increase in size and complexity, ensuring secure, flexible, and privacy-preserving vehicle-to-infrastructure (V2I) authentication remains a major challenge. Existing protocols often focus solely on identity verification, overlooking the need for access control based on vehicle attributes. Furthermore, vehicles must obtain authentication credentials from various trusted entities, including automakers, regulators, and government agencies. However, the absence of a unified credential issuance mechanism introduces fragmentation and inconsistencies during the registration process. To address these issues, we propose a V2I authentication protocol, called PriV2I, that integrates distributed credential issuance, attribute-based access control, and strong anonymity guarantees. During vehicle registration, our approach uses Shamir’s Secret Sharing with a threshold t of n across multiple certification authorities (CAs) to consolidate credentials. A vehicle credential can only be issued by a predefined threshold number of CAs, enhancing security and flexibility. Within the authentication protocol, Pointcheval-Sanders (PS) signatures enable fine-grained access control based on vehicle attributes such as type and role. Meanwhile, noninteractive zero-knowledge proofs protect identity privacy by allowing vehicles to prove credential possession and policy compliance without revealing sensitive information. The proposed scheme also supports batch authentication at Roadside Units (RSUs) to efficiently handle high-density environments and includes a comprehensive revocation mechanism to trace and revoke malicious vehicles promptly and securely. In our implementation, the computation cost during the authentication phase is 75.58 ms. The communication overhead per authentication exchange is 992 bytes across two messages. Overall, the protocol provides a secure, scalable, and privacy-preserving solution tailored to modern VANET environments.

Instruction tuning has emerged as an effective approach that notably improves large language models (LLMs) performance, showing particular promise in natural language generation tasks by producing more diverse, coherent, and task-relevant outputs. However, extending instruction tuning to natural language understanding (NLU) tasks presents significant challenges, primarily due to the difficulty in achieving high-precision responses and the scarcity of large-scale, high-quality instruction data necessary for effective tuning. In this work, we introduce Adversarial Noisy Instruction Tuning (ANIT) to improve NLU performance on LLMs. First, we leverage low-resource techniques to construct noisy instruction datasets. Second, we employ semantic distortion-aware techniques to quantify the intensity of noise within these instructions. Last, we devise an adversarial training method that incorporates a noise response strategy to achieve noisy instruction tuning. ANIT enhances LLMs capability to detect and accommodate semantic distortions in noisy instructions, thereby augmenting their comprehension of task objectives and ability to generate more accurate responses. We evaluate our approach across diverse noisy instructions and semantic distortion quantification methods on multiple NLU tasks. Comprehensive empirical results demonstrate that our method consistently outperforms existing approaches across various experimental settings.

In-Context Learning (ICL) has become an effective paradigm for biomedical named entity recognition (BioNER), allowing large language models to recognize entities using only a few examples without extensive fine-tuning. However, ICL's effectiveness depends heavily on selecting appropriate examples, and current approaches that rely primarily on sentence-level semantic similarity often miss the fine-grained entity distinctions and domain-specific characteristics crucial for BioNER tasks. To address these limitations, we propose Entity-aware In-Context Learning (EICL), a framework that improves example selection by incorporating both entity-centric information and domainspecific terminology overlap. EICL employs an Entity Expansion Module that uses LLM capabilities to create explicit entity representations for better structural alignment, along with a Domain Vocabulary Overlap Retrieval Module that measures domain compatibility through terminological analysis. These components are integrated within a hybrid retrieval strategy that combines semantic similarity with terminology overlap scores for more accurate example selection. Experiments across multiple BioNER benchmark datasets show that EICL outperforms existing ICL example selection methods, demonstrating improved generalization and practical effectiveness for biomedical applications.

Large Language Models (LLMs) have demonstrated significant advancements in various fields, notably in Role-Playing Conversational Agents (RPCAs).However, when confronted with role-specific professional inquiries, LLMsbased RPCAs tend to underperform due to their excessive emphasis on the conversational abilities of characters rather than effectively invoking and integrating relevant expert knowledge.This often results in inaccurate responses.We refer to this phenomenon as the "Knowledge Misalignment" which underscores the limitations of RPCAs in integrating expert knowledge.To mitigate this issue, we have introduced an Anchoring-Guidance Fine-Tuning (AnGFT) Framework into the RPCAs' training process.This involves initially linking the Anchoring-Based System Prompt (ASP) with the LLM's relevant expert domains through diverse prompt construction strategies and supervised fine-tuning (SFT).Following the roleplay enriched SFT, the integration of ASP enables LLMs to better associate with relevant expert knowledge, thus enhancing their response capabilities in role-specific expert domains.Moreover, we have developed four comprehensive metrics-helpfulness, thoroughness, credibility, and feasibility-to evaluate the proficiency of RPCAs in responding to professional questions.Our method was tested across four professional fields, and the experimental outcomes suggest that the proposed AnGFT Framework substantially improves the RPCAs' performance in handling role-specific professional queries, while preserving their robust role-playing abilities.

Large language models (LLMs) have significantly impacted the field of natural language processing (NLP). However, due to the limited domain specificity of the training data and the model’s constrained ability to generalize across complex biomedical data, LLMs continue to encounter challenges related to prediction bias and low generalization in biomedical named entity recognition (BioNER). In this work, we set out to improve the recognition and generalization capabilities of LLMs in BioNER through an Adversarial Selective Training (AST) method. Our method maximizes the adversarial loss to obtain the importance ranking of weights, which guides the model to selectively train to generate counterfactual examples. This strategy aims to force the model to explore the amount of information in the latent space to extract entities, thereby improving the performance of BioNER. Specifically, we conduct in-distribution experiments on five biomedical datasets and out-of-distribution experiments on two datasets. Experimental results show that our method outperforms other LLMs-based methods and significantly improves the performance of BioNER.

Abstract The spatial and temporal uncertainty caused by large propagation delays is a fundamental feature of Underwater Acoustic Networks (UWANs), which seriously affects the performance of the UWANs and also brings challenges to the design of MAC protocols. In this paper, we develop an adaptive MAC protocol based on deep reinforcement learning for UWANs, called ARL-MAC protocol, to intelligently allocate time slots for nodes. Firstly, we design a reward mechanism based on the idea of Time-Domain Interference Alignment (TDIA). We determine the reward according to the combination of the node action and the feedback corresponding to the action. Then, we propose a flexible training mechanism to deal with the ever-changing underwater environment, which improves the fairness of time slot allocation. In addition, we introduce the Deep Recurrent Q-Network (DRQN) algorithm to solve the partially observable information issue. Finally, we evaluate the ARL-MAC protocol with the different number of nodes and changing network environment. Simulation results reveal that the ARL-MAC protocol outperforms other MAC protocols for UWANs in terms of throughput, collision rate and service fairness.

Joint entity and relation extraction (RE) construct a framework for unifying entity recognition and relationship extraction, and the approach can exploit the dependencies between the two tasks to improve the performance of the task. However, the existing tasks still have the following two problems. First, when the model extracts entity information, the boundary is blurred. Secondly, there are mostly implicit interactions between modules, that is, the interactive information is hidden inside the model, and the implicit interactions are often insufficient in the degree of interaction and lack of interpretability. To this end, this study proposes a joint entity and relation extraction model (ESEI) based on E fficient S ampling and E xplicit I nteraction. We innovatively divide negative samples into sentences based on whether they overlap with positive samples, which improves the model’s ability to extract entity word boundary information by controlling the sampling ratio. In order to increase the explicit interaction ability between the models, we introduce a heterogeneous graph neural network (GNN) into the model, which will serve as a bridge linking the entity recognition module and the relation extraction module, and enhance the interaction between the modules through information transfer. Our method substantially improves the model’s discriminative power on entity extraction tasks and enhances the interaction between relation extraction tasks and entity extraction tasks. Experiments show that the method is effective, we validate our method on four datasets, and for joint entity and relation extraction, our model improves the F1 score on multiple datasets.

Finding a desirable sampling estimator has a profound impact on the development of static word embedding models, such as continue-bag-of-words (CBOW) and skip gram (SG), which have been generally accepted as popular low-resource algorithms to generate task-agnostic word representations. Due to the prevalence of large-scale pretrained models, less attention has been paid to these static models in the recent years. However, compared with the dynamic embedding models (e.g., BERT), these static models are straightforward to interpret, cost effective to train, and out-of-box to deploy, thus are still widely used in various downstream models until now. Therefore, it is still of considerable significance to study and improve them, especially the crucial components shared by these static models. In this article, we focus on negative sampling (NS), a key component shared by the sampling-based static models, by investigating and mitigating some critical problems of the sampling core. Concretely, we propose Seeds, a sampling enhanced embedding framework, to learn static word embeddings by a new algorithmic innovation for replacing the NS estimator, in which multifactor global priors are considered dynamically for different training pairs. Then, we implement this framework by four concrete models. For the first two implementations, namely CBOW-GP and SG-GP, both negative words and positive auxiliaries are sampled. And for the other two implementations, CBOW-GN and SG-GN, estimations are simplified by sampling only the negative instances. Extensive experimental results across a variety of standard intrinsic and extrinsic tasks demonstrate that embeddings learned by the proposed models outperform their NS-based counterparts, such as CBOW-NS and SG-NS, as well as other strong baselines.

The dramatic proliferation on emerging Internet-of-Things (IoT) makes our telecommunications networks more and more congested. Due to the flexible deployment and spectrum supplement capabilities, cognitive radio based unmanned aerial vehicles (CUAVs) have been regarded as a promising solution to help the network offload the overwhelming traffic. For the CUAV-assisted network, how to offload as much traffic as possible is significant. It is necessary to jointly consider both sides on data collection and data transmission, which, however, is a very challenging problem due to the heterogeneous and uncertain environment on both traffic demand and spectrum availability. In this paper, aiming at maximizing the offloaded traffic, we propose a joint strategy on trajectory design, time division, and spectrum access. Considering the unobtainable environmental information on both traffic demand and spectrum availability, we further develop a model-free deep reinforcement learning (DRL) based solution for the T2S joint strategy, so that the CUAV could make the best decisions autonomously under the uncertain environment. Simulation results have shown the effectiveness of the designed DRL solution and also the offloading efficiency of the proposed T2S strategy.

Participatory sensing is often used in environmental or personal data monitoring, wherein a number of participants collect data using their mobile intelligent devices for earning the incentives. However, a lot of additional information is submitted along with the data, such as the participant’s location, IP and incentives. This multimodal information implicitly links to the participant’s identity and exposes the participant’s privacy. In order to solve the issue of these multimodal information associating with participants’ identities, this paper proposes a protocol to ensure anonymous data reporting while providing a dynamic incentive mechanism simultaneously. The proposed protocol first establishes a submission schedule by anonymously selecting a slot in a vector by each member where every member and system entities are oblivious of other members’ slots and then uses this schedule to submit the all members’ data in an encoded vector through bulk transfer and multiplayer dining cryptographers networks (DC-nets) . Hence, the link between the data and the member’s identity is broken. The incentive mechanism uses blind signature to anonymously mark the price and complete the micropayments transfer. Finally, the theoretical analysis of the protocol proves the anonymity, integrity, and efficiency of this protocol. We implemented and tested the protocol on Android phones. The experiment results show that the protocol is efficient for low latency tolerable applications, which is the cases with most participatory sensing applications, and they also show the advantage of our optimization over similar anonymous data reporting protocols.

Simple linear models, which usually learn word-level representations that are later combined to form document representations, have recently shown impressive performance. To improve the performance of document-level classification, it is crucial to explore the factors affecting the quality of the document vector. In this paper, we propose the concept of containers and further explore the properties of word containers and document containers by experiments and theoretical demonstrations. We find that the document container has a fixed capacity and that the document vector obtained by a simple average of too many word embeddings undoubtedly cannot be fully loaded by the container and will lose some semantic and syntactic information on very large text datasets. We also propose an efficient approach for document representation, using clustering algorithms to divide a document container into several subcontainers and establishing the relationship between the subcontainers. We additionally report and discuss the properties of two methods of clustering algorithms, DVEM-Kmeans and DVEM-Random, on large text datasets by sentiment analysis and topic classification tasks. Compared to simple linear models, the results show that our models outperform the existing state-of-the-art in generating high-quality document representations for document-level classification relatedness tasks. Our approaches can also be introduced to other models based on neural networks, such as convolutional neural networks, recurrent neural networks and generative adversarial networks, in supervised or semisupervised settings.

Dynamic opportunistic channel access with software‐defined radio at a network layer in distributed cognitive IoT introduces a concurrent channel selection along with end‐to‐end route selection for application data transmission. State‐of‐the‐art cognitive IoT big data‐based routing protocols are not explored in terms of how the spectrum management is being coordinated with the network layer for concurrent channel route selection during end‐to‐end channel route discovery for data transmission of IoT and big data applications. In this paper, a reactive big data‐based “cognitive dynamic source routing protocol” is proposed for cognitive‐based IoT networks to concurrently select the channel route at the network layer from source to destination. Experimental results show that the proposed protocol cognitive DSR with concurrent channel route selection criteria is outperformed. This will happen when it is compared with the existing distributed cognitive DSR with independent channel route application data transmission.

Abstract Key management is an important security service for protecting wireless sensor networks (WSNs). Among various existing schemes, exclusion basis system (EBS) is a practical solution that can be easily implemented to provide long‐term and flexible protection for WSNs. The involved rekeying strategy in EBS can efficiently evict the compromised node and update the key system. However, the relatively small key pool leads to high correlation among the generated key rings. Consequently, it is almost impossible for EBS‐based schemes to efficiently defend collusion attack with their rekeying mechanisms. In this paper, we first analyze the impact of collusion attack on WSNs, especially that in the case where the keys of the compromised nodes can form a connected graph. Then, we propose a novel key management scheme based on EBS. The proposed scheme is termed ast‐EEBS because it can effectively resist the referred collusion attack formed byt(t > 1) nodes. Furthermore, we assume that the proposed scheme is implemented in hierarchical WSNs. In this case, two layers oft‐EEBS administrative keys are used. The upper layer is implemented among the base station and all cluster leaders, while the lower layer involves at‐EEBS for each cluster. The results of performance evaluation show that the proposed scheme has better resistance to collusion attack than other schemes. Therefore, the proposed scheme can provide better security service for WSNs. Copyright © 2017 John Wiley & Sons, Ltd.