Zhezheng Hao, Hong Wang, Jian Luo, Jianqing Zhang, Yuyan Zhou, Qiang Lin, Can Wang, Hande Dong, Jiawei Chen. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

This paper proposes a new hierarchically tunable six-dimensional movable antenna (HT-6DMA) architecture for base station (BS) in future wireless networks, aiming to improve the performance of both wireless communication and sensing. The HT-6DMA BS consists of multiple antenna arrays that can flexibly move on a spherical surface, with their three-dimensional (3D) positions and 3D rotations/orientations efficiently characterized in the global spherical coordinate system (SCS) and their individual local SCSs, respectively. As a result, the 6DMA system is hierarchically tunable in the sense that each array’s global position and local rotation can be separately adjusted in a sequential manner with the other being fixed, thus greatly reducing their design complexity and improving the achievable performance. In particular, we consider an HT-6DMA BS serving multiple single-antenna users in the uplink communication or sensing potential unmanned aerial vehicles (UAVs)/drones in a given airway area. Specifically, for the communication scenario, we aim to maximize the average sum rate of communication users in the long term by optimizing the positions and rotations of all 6DMA arrays at the BS. For the airway sensing scenario, we maximize the minimum received sensing signal power along the airway by optimizing the 6DMA arrays’ positions and rotations along with the BS’s transmit covariance matrix. Despite that the formulated problems are both non-convex and challenging to solve, we propose efficient solutions to them by exploiting the hierarchical tunability of positions/rotations of 6DMA arrays in our proposed model. Numerical results show that the proposed HT-6DMA design significantly outperforms not only the traditional BS with fixed-position antennas (FPAs), but also the existing 6DMA scheme based on alternating array position/rotation optimization. Furthermore, it is unveiled that the performance gains of HT-6DMA mostly come from the arrays’ global position adjustments on the spherical surface, rather than their local rotation adjustments, which provides a useful guide for implementing 6DMA systems under practical performance-complexity trade-off consideration.

Recent studies showed that the generalization of neural networks is correlated with the sharpness of the loss landscape and flat minima suggests a better generalization ability than sharp minima. In this paper, we propose a novel method called optimum shifting, which changes the parameters of a neural network from a sharp minimum to a flatter one while maintaining the same training loss value. Our method is based on the observation that when the input and output of a neural network are fixed, the matrix multiplications within the network can be treated as systems of under-determined linear equations, enabling adjustment of parameters in the solution space, which can be simply accomplished by solving a constrained optimization problem. Furthermore, we introduce a practical stochastic optimum shifting technique utilizing the neural collapse theory to reduce computational costs and provide more degrees of freedom for optimum shifting. Extensive experiments with various deep neural network architectures on benchmark datasets demonstrate the effectiveness of our method.

Deep learning methods have demonstrated state-of-the-art performance in image restoration, especially when trained on large-scale paired datasets. However, acquiring paired data in real-world scenarios poses a significant challenge. Unsupervised restoration approaches based on generative adversarial networks (GANs) offer a promising solution without requiring paired datasets. Yet, these GAN-based approaches struggle to surpass the performance of conventional unsupervised GAN-based frameworks without significantly modifying model structures or increasing the computational complexity. To address these issues, we propose a self-collaboration (SC) strategy for existing restoration models. This strategy utilizes information from the previous stage as feedback to guide subsequent stages, achieving significant performance improvement without increasing the framework's inference complexity. The SC strategy comprises a prompt learning (PL) module and a restorer ($Res$Res). It iteratively replaces the previous less powerful fixed restorer $\overline{Res}$Res¯ in the PL module with a more powerful $Res$Res. The enhanced PL module generates better pseudo-degraded/clean image pairs, leading to a more powerful $Res$Res for the next iteration. Our SC can significantly improve the $Res$Res 's performance by over 1.5 dB without adding extra parameters or computational complexity during inference. Meanwhile, existing self-ensemble (SE) and our SC strategies enhance the performance of pre-trained restorers from different perspectives. As SE increases computational complexity during inference, we propose a re-boosting module to the SC (Reb-SC) to improve the SC strategy further by incorporating SE into SC without increasing inference time. This approach further enhances the restorer's performance by approximately 0.3 dB. Additionally, we present a baseline framework that includes parallel generative adversarial branches with complementary "self-synthesis" and "unpaired-synthesis" constraints, ensuring the effectiveness of the training framework. Extensive experimental results on restoration tasks demonstrate that the proposed model performs favorably against existing state-of-the-art unsupervised restoration methods.

When taking images against strong light sources, the resulting images often contain heterogeneous flare artifacts. These artifacts can significantly affect image visual quality and downstream computer vision tasks. While collecting real data pairs of flare-corrupted/flare-free images for training flare removal models is challenging, current methods utilize the direct-add approach to synthesize training data. However, these methods do not consider automatic exposure and tone mapping in the image signal processing pipeline (ISP), leading to the limited generalization capability of deep model training using such data. Besides, existing light source recovery methods hardly recover multiple light sources due to the different sizes, shapes, and illuminance of various light sources. In this paper, we propose a solution to improve the performance of lens flare removal by revisiting the ISP, remodeling the principle of automatic exposure in the synthesis pipeline, and designing a more reliable light source recovery strategy. The new pipeline approaches realistic imaging by discriminating the local and global illumination through a convex combination, avoiding global illumination shifting and local over-saturation. Moreover, the current deep models are only generalized to specific devices due to the diversity of cameras' ISPs. To achieve better generalization on different devices, we formulate the generalization problem as an adversarial training problem and embed an adversarial curve learning (ACL) paradigm in the synthesis pipeline to gain better performance. For recovering multiple light sources, our strategy convexly averages the input and output of the neural network based on illuminance levels, thereby avoiding the need for a hard threshold in identifying light sources. We also contribute a new flare removal testing dataset containing the flare-corrupted images captured by fifteen types of consumer electronics. The dataset facilitates the verification of the generalization capability of flare removal methods. Extensive experiments show that our solution can effectively improve the performance of lens flare removal and push the frontier toward more general situations.

We introduce a web-based SQL query visualisation tool aimed at improving students' understanding of SQL query execution. The tool, available as an Open Educational Resource (OER), allows users to visualise the stages of a SQL query on an uploaded SQLite database. The web interface enables students to interactively explore each stage of the query and view results in a dynamically updated table. This initial version of the tool is intended to inspire further development and enhancement while providing educators with a valuable resource for teaching SQL concepts.

The advent of large language models (LLMs) like GPT-4 has catalyzed the exploration of multi-task learning (MTL), in which a single model demonstrates proficiency across diverse tasks.Task arithmetic has emerged as a costeffective approach for MTL.It enables performance enhancement across multiple tasks by adding their corresponding task vectors to a pre-trained model.However, the current lack of a method that can achieve optimal performance with low computational cost and protecting the data privacy, which limits their application to LLMs.In this paper, we propose Model Exclusive Task Arithmetic for merging GPT-scale models (MetaGPT), which formalizes the objective of model merging into a multi-task learning framework, aiming to minimize the average loss difference between the merged model and each individual task model.Since data privacy limits the use of multi-task training data, we leverage LLMs' local linearity and task vectors' orthogonality to separate the data term and scaling coefficients term and derive a model-exclusive task arithmetic method.Our proposed MetaGPT is dataagnostic and bypasses the heavy search process, making it cost-effective and easy to implement for LLMs.Extensive experiments demonstrate that MetaGPT leads to improvements in task arithmetic and achieves state-of-the-art performance on multiple tasks.

Machine learning (ML) approaches have been successfully applied to accelerating exact combinatorial optimization (CO) solvers. However, many of them fail to explain what patterns they have learned that accelerate the CO algorithms due to the black-box nature of ML models like neural networks, and thus they prevent researchers from further understanding the tasks they are interested in. To tackle this problem, we propose the first graph-based algorithm discovery framework—namely, graph symbolic discovery for exact combinatorial optimization solver (GS4CO)—that learns interpretable branching policies directly from the general bipartite graph representation of CO problems. Specifically, we design a unified representation for symbolic policies with graph inputs, and then we employ a Transformer with multiple tree-structural encodings to generate symbolic trees end-to-end, which effectively reduces the cumulative error from iteratively distilling graph neural networks. Experiments show that GS4CO learned interpretable and lightweight policies outperform all the baselines on CPU machines, including both the human-designed and the learning-based. GS4CO shows an encouraging step towards general algorithm discovery on modern CO solvers.

Reconfigurable intelligent surface (RIS) has been envisioned as a highly promising technology for future wireless communication networks. Very recently, a novel beyond diagonal (BD)-RIS architecture has been proposed. This new architecture remarkably extends the traditional diagonal RIS model and yields much more powerful beamforming capability. Meanwhile, however, the emerging symmetry and orthogonality conditions imposed onto BD-RIS’ reflection matrix make its optimization highly difficult, especially when BD-RIS must satisfy numerous additional constraints. This difficulty arises in many BD-RIS applications and has remained unsolved so far. To resolve the above challenge, leveraging the penalty dual decomposition methodology, this paper proposes a novel unified approach that can optimize BD-RIS configuration when it is involved in any number of nonconvex constraints. Especially, we utilize our new approach to solve the power minimization and energy efficiency maximization problems when BD-RIS involves multiple quality-of-service constraints, which have not yet been solved in the literature. Besides, our new approach can also efficiently solve the sum-rate maximization in the BD-RIS assisted system by providing a new analytic-update-based solution, which is more efficient than existing methods. Extensive numerical results demonstrate the effectiveness of our new approach and the significant benefit of BD-RIS over the conventional diagonal RIS.

Despite its high energy and hardware efficiency, some defects of the reconfigurable intelligence surface (RIS) technology have come to be realized, including the severe fading loss and restricted-to-half-space coverage. This paper proposes a novel double-faced-active (DFA)-RIS structure to overcome these defects. Besides, we utilize this novel DFA-RIS to improve power saving of the communication system. Unlike traditional power saving literature, we aim at fulfilling queueing stability and long-term power minimization in a downlink system assisted by the DFA-RIS, with a realistic data arriving process taken into consideration. Enlightened by Lyapunov control theory, we propose an online optimization strategy that adaptively adjusts DFA-RIS configuration. Each online problem can be efficiently solved by leveraging alternative directional method of multipliers (ADMM) method. Numerical results demonstrate the effectiveness of our proposed Lyapunov-guided strategy and DFA-RIS’ superiority over the classical passive RIS.

When taking images against strong light sources, the resulting images often contain heterogeneous flare artifacts. These artifacts can importantly affect image visual quality and downstream computer vision tasks. While collecting real data pairs of flare-corrupted/flare-free images for training flare removal models is challenging, current methods utilize the direct-add approach to synthesize data. However, these methods do not consider automatic exposure and tone mapping in image signal processing pipeline (ISP), leading to the limited generalization capability of deep models training using such data. Besides, existing methods struggle to handle multiple light sources due to the different sizes, shapes and illuminance of various light sources. In this paper, we propose a solution to improve the performance of lens flare removal by revisiting the ISP and remodeling the principle of automatic exposure in the synthesis pipeline and design a more reliable light sources recovery strategy. The new pipeline approaches realistic imaging by discriminating the local and global illumination through convex combination, avoiding global illumination shifting and local over-saturation. Our strategy for recovering multiple light sources convexly averages the input and output of the neural network based on illuminance levels, thereby avoiding the need for a hard threshold in identifying light sources. We also contribute a new flare removal testing dataset containing the flare-corrupted images captured by ten types of consumer electronics. The dataset facilitates the verification of the generalization capability of flare removal methods. Extensive experiments show that our solution can effectively improve the performance of lens flare removal and push the frontier toward more general situations.

Although reconfigurable intelligent surface (RIS) technology has manifested great potentials in improving wireless network’s power saving, most existing literature restricts to pure physical (PHY) layer beamforming design and neglects the impact of media access control (MAC) layer’s data traffic flows. Simultaneously, current RIS technology suffers from defects — the severe fading loss and the limitation of half-space coverage. This paper aims to perform a cross-layer design via jointly optimizing MAC layer scheduling and PHY layer RIS beamforming to reduce power consumption. Besides, we propose a novel double-faced-active (DFA)-RIS architecture to promote RIS’ capability. The proposed design task leads to a highly challenging stochastic problem to minimize long-term power consumption while stabilizing queues. Inspired by Lyapunov control theory, we propose an online optimization strategy to resolve this challenge. Via exploiting alternative directional method of multipliers (ADMM), we develop an analytic-based solution to solve the online sub-problems highly efficiently without resorting to any numerical solvers. Our strategy theoretically guarantees all queues’ stability and achieves a tunable trade-off between the power expenditure and queue lengths. Extensive numerical results are presented to demonstrate the effectiveness of our proposed cross-layer design and the DFA-RIS’ advantage over other cutting-the-edge RIS architectures.