Hong Huang, Decheng Wu, Qiangqiang Hu, Guanghua Yu, Jinhai Yang, Jianchen Zhu, Xue Liu, Dapeng Wu. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Siheng Li, Kejiao Li, Zenan Xu, Guanhua Huang, Kun Li, Haoyuan Wu, Wujiajia, Zihao Zheng, Chenchen Zhang, Kun Shi, Xue Gong, Qi Yi, Ruibin Xiong, Tingqiang Xu, Yuhao Jiang, Jianfeng Yan, Yuyuan Zeng, Guanghui Xu, Jinbao Xue, Zhijiang xu, Zheng Fang, Shuai LI, Qibin Liu, Xiaoxue Li, Zhuoyu Li, Yangyu Tao, Fei Gao, Cheng Jiang, Bochao Wang, Kai Liu, Jianchen Zhu, Wai Lam, Bo Zhou, Di Wang. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Sustainable energy supply and high-speed communications are two significant needs for the upcoming 6G applications. This paper introduces a self-aligning resonant beam system for simultaneous light information and power transfer (SLIPT), employing a novel coupled spatially distributed resonator (CSDR). The system utilizes a resonant beam for efficient power delivery and a second-harmonic beam for concurrent data transmission, inherently minimizing echo interference and enabling bidirectional communication. Through comprehensive analyses, we investigate the CSDR’s stable region, beam evolution, and power characteristics in relation to working distance and device parameters. Numerical simulations validate the CSDR-SLIPT system’s feasibility by identifying a stable beam waist location for achieving accurate mode-match coupling between two spatially distributed resonant cavities and demonstrating its operational range and efficient power delivery across varying distances. The research reveals the system’s benefits in terms of both safety and energy transmission efficiency. We also demonstrate the trade-off among the reflectivities of the cavity mirrors in the CSDR. Besides, an experiment was conducted to verified the feasibility of self-aligning beam generation and safety under the designed structure. These findings offer valuable design insights for resonant beam systems, advancing SLIPT with significant potential for remote device connectivity.

Speculative decoding (SD), where an extra draft model is employed to provide multiple **draft** tokens first and then the original target model verifies these tokens in parallel, has shown great power for LLM inference acceleration. However, existing SD methods suffer from the mutual waiting problem, i.e., the target model gets stuck when the draft model is *guessing* tokens, and vice versa. This problem is directly incurred by the asynchronous execution of the draft model and the target model, and is exacerbated due to the fixed draft length in speculative decoding. To address these challenges, we propose a conceptually simple, flexible, and general framework to boost speculative decoding, namely **P**arallel sp**E**culative decoding with **A**daptive d**R**aft **L**ength (PEARL). Specifically, PEARL proposes *pre-verify* to verify the first draft token in advance during the drafting phase, and *post-verify* to generate more draft tokens during the verification phase. PEARL parallels the drafting phase and the verification phase via applying the two strategies, and achieves adaptive draft length for different scenarios, which effectively alleviates the mutual waiting problem. Experiments on various text generation benchmarks demonstrate the effectiveness of our PEARL, leading to a superior speedup performance up to **4.43$\times$** and **1.50$\times$**, compared to auto-regressive decoding and vanilla speculative decoding, respectively.

As a recently proposed reconfigurable intelligent surface (RIS) architecture, active RIS has drawn considerable interest. The important feature of the active RIS is its ability to strengthen the impinging signals to mitigate the multiplicative fading effect inherent in passive RIS-aided systems. Herein, we explore the performance of an active RIS-aided uplink multi-user massive multiple-input multiple-output (MIMO) system, considering the phase noise at the RIS. Furthermore, a two-timescale scheme is utilized, where the base station (BS) beamforming is designed based on the instantaneous aggregated channel state information (CSI), while the statistical CSI is used for designing the phase shifts of the active RIS. In addition, the linear minimum mean square error (LMMSE) estimator is adopted to estimate the aggregated channel, which combines both the cascaded and direct channels. According to the estimated channel, a closedform expression for the lower bound of achievable rate is derived. Based on the theoretical expressions, the power scaling laws for the considered system are also investigated. Specifically, when each user’s transmit power is proportionally reduced by the quantity of BS antennasMor RIS elementsN, we find that the amplified thermal noise causes the lower bound of the achievable rate to approach zero asMorNtends to infinity. Moreover, an optimization approach based on a genetic algorithm (GA) is introduced to obtain the optimal phase shifts for maximizing the achievable rate. Numerical results reveal that the active RIS can greatly enhance the performance of the massive MIMO system compared to its passive counterpart.

Large language models (LLMs) have proven to be very superior to conventional methods in various tasks.However, their expensive computations and high memory requirements are prohibitive for deployment.Model quantization is an effective method for reducing this overhead.The problem is that in most previous works, the quantized model was calibrated using a few samples from the training data, which might affect the generalization of the quantized LLMs to unknown cases and tasks.Hence in this work, we explore an important question: Can we design a data-free quantization method for LLMs to guarantee its generalization performance?In this work, we propose EasyQuant, a trainingfree and data-free weight-only quantization algorithm for LLMs.Our observation indicates that two factors: outliers in the weight and quantization ranges, are essential for reducing the quantization error.Therefore, in EasyQuant, we leave the outliers (less than 1%) unchanged and optimize the quantization range to reduce the reconstruction error.With these methods, we surprisingly find that EasyQuant achieves comparable performance to the original model.Since EasyQuant does not depend on any training data, the generalization performance of quantized LLMs are safely guaranteed.Moreover, EasyQuant can be implemented in parallel so that the quantized model could be attained in a few minutes even for LLMs over 100B.To our best knowledge, we are the first work that achieves comparable performance with datadependent algorithms under a data-free setting and our algorithm runs over 10 times faster than the data-dependent methods.

The coded aperture snapshot spectral imager (CASSI) system senses spatial and spectral information using a binary coded aperture and a dispersive element, thus the quality of reconstructed hyperspectral images is mainly determined by the structure of coded apertures. Traditional coded apertures (Random, Bernoulli, etc.), encoding hyperspectral images in focal array plane, suffer from suboptimal reconstruction accuracy. Therefore, optimizing coded aperture design improves the reconstruction quality for the scene. In this paper, a fast iterative algorithm performing coded apertures is investigated. Since calculating restricted isometry constant is NP-hard, the structure of coded apertures is alternatively designed based on mutual coherence property, via CASSI sensing matrice improvement. It has been proven that by exploiting blue noise spatial-spectral characteristics, the clusters of one-valued (white) entries in the coded aperture ensembles can be reduced. Finally, simulations are carried out on a real data set, showing the superiority of the proposed coded apertures in terms of various evaluation metrics.

This paper aims at developing a clustering approach with spectral images directly from the compressive measurements of coded aperture snapshot spectral imager (CASSI). Assuming that compressed measurements often lie approximately in low dimensional subspaces corresponding to multiple classes, state of the art methods generally obtains optimal solution for each step separately but cannot guarantee that it will achieve the globally optimal clustering results. In this paper, a low-rank subspace representation (LRSR) algorithm is proposed to perform clustering on the compressed measurements. In addition, a subspace structured norm is added into the objective of low-rank representation problem exploiting the fact that each point in a union of subspaces can be expressed as a sparse linear combination of all other points and that the matrix of the points within each subspace is low rank. Simulation with real dataset illustrates the accuracy of the proposed spectral image clustering approach.