Zhen Yang, Wei Du, Jie Wang, Wenze Zhou, Xiangfeng Meng, Zhengyang Wang, Suping Sun, Ziwei Du, Haodong Zou, Jie Chen, Yongbin Liu, Shicheng Tan, Jiahao Ying, Shu Zhao. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Knowledge Base Question Answering (KBQA) leverages structured knowledge bases to offer superior interpretability and hallucination resistance, making it a critical technology for precise knowledge reasoning.However, the prevailing LLM-based generate-then-execute formulation of semantic parsing is limited by strict syntactic constraints, making it primarily prone to structural deviations that render queries unexecutable, while suffering from semantic deviations that yield incorrect execution results.To address these challenges, we propose the Execution as Verification (EVER) framework, reframing semantic parsing as an iterative, self-correcting reasoning process driven by execution feedback.First, motivated by the insight that query executability serves as a strong proxy for answer correctness, we introduce Fine-Grained Execution-Aware Planning.This mechanism decomposes complex semantic parsing into a sequence of stepwise reasoning processes oriented by executability verification, ensuring high query executability.We further design a Self-Guided Semantic Correction mechanism based on execution result verification, utilizing execution feedback to verify and calibrate semantic deviations, thereby ensuring the semantic correctness of executable queries.Experimental results on the WebQSP and CWQ datasets demonstrate that our method achieves significant improvements in both query executability and answer accuracy, achieving stateof-the-art performance, particularly in complex multi-hop scenarios.Our code is available at https://github.com/ahu-zmh/EVER.

Large Language Models (LLMs) have achieved remarkable performance across various artificial intelligence applications. However, current LLMs cannot be deployed directly on edge nodes due to their large number of parameters. Fortunately, model compression technology has been proposed to reduce the computational workload and memory usage of LLMs, enabling further edge-based LLM services. However, existing research typically concentrates on isolated compression algorithms and lacks a comprehensive perspective on how to leverage these techniques for practical, end-to-end LLM deployment in edge environments. In this survey, we review edge-oriented LLM compression techniques and software–hardware co-design strategies to enable efficient LLM deployment on resource-constrained edge systems and guide future research in this area. First, we analyze techniques for LLM compression from the perspective of cloud–edge collaborative intelligence, including model quantization, parameter pruning, and knowledge distillation. Second, we present several hybrid model frameworks tailored to dynamic, heterogeneous edge environments, based on model architecture, application scenarios, and combination selection. Third, we further refine a four-layer software–hardware codesign and an overhead-aware LLM deployment optimization. Finally, we discuss the challenges of current model compression approaches and offer insights into future research directions, with a focus on edge-based LLM services.

Person detection in videos is vital for area admission and public safety. Existing studies have made significant progress in improving the accuracy of this task on the cloud. Meanwhile, with people's increasing awareness of privacy protection, there is a surging demand for privacy not being transmitted and processed by the cloud. Thus, providing services on edges becomes a promising solution. The dilemma is that edges are typically resource-constrained and cannot support the deployment of large models. However, tiny models that fit resource-constrained edges generally have unsatisfactory performance in accuracy and efficiency. To this end, we propose a Logical Correction Enabled Collaborative Person Detection Inference (LC-CPDI) framework for resource-constrained edges. First, we formulate the problem studied with a delay minimization objective. Second, we design a logical correction scheme to perceive abnormal predictions and perform corrections to improve accuracy. Third, a hybrid position prediction algorithm is proposed to replace time-consuming inference for simple scenarios. Finally, we design a collaborative inference scheme that enables frame outsourcing to idle edges to reduce the inference delay. We implemented LC-CPDI on a testbed designed with commercial edges. The experiments on real-world datasets show the effectiveness of LC-CPDI with up to 41.8% delay reduction on average and near 2% recall improvement.

To support delay-critical applications, migrating services from the cloud to edge servers (ESs) can effectively reduce service delay. However, in such a resource-constrained scenario, providing heterogeneous services to meet diverse user needs poses significant challenges. Specifically, ESs have limited computational resources and are often energy-constrained, complicating service placement and provisioning. Existing studies propose collaboration schemes to improve resource utilization and reduce service delay. Nevertheless, these works heavily depend on full-service coverage by the cloud or rely on coarse-grained (e.g., time-cycle level) strategies that inevitably waste energy in idle time slots, which are unsuitable for energy-constrained settings and dynamic edge environments. To address these challenges, we propose a novel edge-edge collaboration method tailored for heterogeneous edge service provisioning in energy-constrained networks. First, we formulate the heterogeneous service provisioning problem with the objective of delay minimization under energy constraints and prove its NP-hardness. Our framework leverages latest task statistics to decide the service lifetime in a fine-grained time slot level, so that enables edge collaboration to maximize resource utilization and adaptability in resource-limited conditions. Specifically, we decompose the problem into three subproblems: service placement, service lifetime decision, and task scheduling, and we design targeted lightweight solutions for each, ensuring low delay and efficient energy usage. Finally, we validate our method through comprehensive simulations on real-world datasets and implement it on a testbed with three ESs. Results demonstrate that our approach reduces service delays by an average of 69.6% across various energy-constrained scenarios.

Delay-critical small object detection is crucial for real-time applications, such as autonomous driving and aerial surveillance. While cloud-based approaches incur substantial transmission latency, edge-based detection can avoid such delays; however, it faces its own challenges: limited computational capacity at the edge often leads to unsatisfactory inference delays and low accuracy when employing lightweight models. To tackle these issues, this paper presents an adaptive image batching and slicing framework for delay-sensitive small object detection in distributed edge environments. We first formulate the problem as a mixed-integer nonlinear programming model, which is NP-hard. The proposed scheme then aggregates tasks with tight deadlines into batches processed locally or cooperatively across the edge network, thus meeting stringent timing constraints. For tasks with relatively relaxed deadlines, we introduce an adaptive image slicing strategy that preserves fine-grained pixel information to boost detection accuracy without violating deadline requirements. The framework is implemented on NVIDIA Jetson edge devices and evaluated using real-world datasets. Experimental results show that our approach outperforms state-of-the-art baselines by an average accuracy gain of 0.5%, while improving the task success rate by 19%.

Collaborative edge computing has been widely advo-cated by network operators and service providers to promote the quality of service (QoS), provisioning diverse delay-sensitive and computation-intensive applications. Existing studies mainly focus on cloud-edge collaboration, since cloud servers have massive resources to provide diverse services and edge servers can provide low-delay services with close proximity to end users. However, in scenarios that capture privacy, e.g., personal bioinformation and business areas, there is a great need for zero cloud involvement. Moreover, current edge servers are typically energy-constrained, which poses great challenges in enabling high-QoS services in ever-densely deployed edge networks. To tackle these issues, in this paper, we study the energy-constrained edge-edge collaboration problem. First, we formulate the edge-edge collaboration with delay minimization and energy reduction aims and prove its NP-hardness. Second, we propose a novel Fine-Grained Service Lifetime Optimization (FGSLO) scheme as a possible solution. The problem is then transformed and decoupled into three sub-problems, namely service placement, service lifetime decision, and task scheduling, which are solved by our proposed method, respectively. Finally, real-world data-driven experimental results show that FGSLO is capable of reducing 21.4%~90.1 % system delay in different energy-constrained scenarios, compared to baselines without service lifetime control.

With the surge of end devices and intelligent services, computing resource has begun to migrate from the cloud to end devices to meet the growing demand of users, forming a hierarchical resource distribution pattern in cloud-edge-end networks. Existing research work focuses on using edge computing technique to build a cloud-edge collaborative service offloading and task scheduling method to reduce service delay or energy consumption. However, different user groups and application scenarios may have different preferences for service quality requirements even for the same kind of service. For example, video analysis in autonomous driving focuses more on delay while video analysis in surveillance focuses more on accuracy. Meanwhile, heterogeneous cloud-edge-end devices have significant differences in the amount of resources, which poses great challenges for efficient provisioning of services. To solve this problem, we intend to propose multi-view service provisioning method in cloud-edge-end networks with hierarchically distributed resources. Firstly, we design a mapping scheme between service quality and heterogeneous resource occupation to estimate the amount of resources required for a given requirement. Secondly, we utilize model compression methods to customize powerful large models into smaller and lighter one according to the requirements of tasks. Thirdly, as resources are distributed hierarchically in cloud-edge-end networks, efficient service placement should be carried out with the goal of achieving diverse needs. The effectiveness of the proposed method is demonstrated through numerical simulations compared to state-of-the-art baselines and experiments on an implemented prototype system.

Collaborative Mobile Edge Computing (MEC) has emerged as a promising approach to provide low service latency for computation-intensive Internet of Things applications, facilitated by the cooperation of edge-edge and edge-cloud resources. However, existing collaborative MEC methods typically restrict the collaborative processing between any two Edge Servers (ESs) or one ES and the cloud server for a task request, limiting the exploitation of available resources on other ESs. Moreover, these conventional methods rely on offline task partitioning, potentially leading to extended make-span, especially when ES computing capacities exhibit heterogeneity. In this paper, we propose an innovative method named SMCoEdge. This method performs dynamic parallel multi-ES selection and workload allocation in heterogeneous collaborative MEC environments, thus simultaneously enabling multiple ESs' idle resources to accelerate task processing. We formulate our problem into an online linear programming problem, with the objective of minimizing task computing and transmission make-spans. To enhance computational efficiency, we decompose the problem into two stages: multi-ES selection and workload allocation. Then, we propose an online Deep Reinforcement Learning based Simultaneous Multi-ES Offloading (DRL-SMO) algorithm along with a top-$k$deep Q-learning network model to effectively solve our problem, where an efficient algorithm is proposed to achieve the optimal solution for the workload allocation stage. Furthermore, we provide a theoretical performance analysis, demonstrating that the DRL-SMO algorithm achieves a near-optimal solution for our problem within an approximate linear time complexity. Finally, our extensive experimental results demonstrate the substantial advantages of our method. It consistently reduces the average make-span by 19.63% and keeps a lower offloading failure rate, when compared to state-of-the-art methods. These findings underline the efficacy of our method in enhancing collaborative MEC performance.

To meet the growing need of mobile traffic with Ultra-Reliable and Low Latency Communication (URLLC) requirement, 5G New Radio (NR) is extending from licensed band to unlicensed band on which Wi-Fi has already been operated, resulting in coexisting NR/Wi-Fi network. Existing works have made great efforts on throughput and latency of coexisting NR/Wi-Fi network. However, excessive NR requests offloaded from licensed band lead to unfair utilization of unlicensed band, which further causes unsatisfaction on URLLC and performance degradation of Wi-Fi. In this paper, we propose a novel Reinforcement Learning based Transmission Revoking Approach (RL-TRA) to address this problem aiming at fairer utilization of unlicensed band restrained by URLLC. Firstly, we formulate the coexistence problem of NR/Wi-Fi as integer non-linear programming and show its NP-hardness. Secondly, we decompose the problem into three sub-problems, namely redundancy determining, request scheduling, and transmission revoking. The former two sub-problems are solved with our proposed method to satisfy URLLC requirement. We further propose a novel transmission revoking mechanism when tackling transmission revoking sub-problem, aiming at maintaining fairness of coexisting NR/Wi-Fi network. Finally, simulation results verify the effectiveness of RL-TRA. By using our method, the fairness is improved by 16.5% averagely with only 1.77% loss on success rate of URLLC requests compared with baselines.