Large language models (LLMs) have advanced to encompass extensive knowledge across diverse domains. Yet controlling what a LLMs should not know is important for ensuring alignment and thus safe use. However, effective unlearning in LLMs is difficult due to the fuzzy boundary between knowledge retention and forgetting. This challenge is exacerbated by entangled parameter spaces from continuous multi-domain training, often resulting in collateral damage, especially under aggressive unlearning strategies. Furthermore, the computational overhead required to optimize State-of-the-Art (SOTA) models with billions of parameters poses an additional barrier. In this work, we present ALTER, a lightweight unlearning framework for LLMs to address both the challenges of knowledge entanglement and unlearning efficiency. ALTER operates through two phases: (I) high entropy tokens are captured and learned via the shared A matrix in LoRA, followed by (II) an asymmetric LoRA architecture that achieves a specified forgetting objective by parameter isolation and unlearning tokens within the target subdomains. Serving as a new research direction for achieving unlearning via token-level isolation in the asymmetric framework. ALTER achieves SOTA performance on TOFU, WMDP, and MUSE benchmarks with over 95% forget quality and shows minimal side effects through preserving foundational tokens. By decoupling unlearning from LLMs' billion-scale parameters, this framework delivers excellent efficiency while preserving over 90% of model utility, exceeding baseline preservation rates of 47.8-83.6%.

Jie Ou, Jinyu Guo, Shiyao Guo, Yuang Li, Ruiqi Wu, Zhaokun Wang, Wenyi Li, Wenhong Tian. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Jingwen Pu, Mingjun Shi, Xinrui Ren, Yizhe Wang, Xinyu Zhang, Zhaokun Wang, Kun She. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Zhaokun Wang, Jinyu Guo, Jingwen Pu, Hongli Pu, Meng Yang, Xunlei Chen, Jie Ou, Wenyi Li, Guangchun Luo, Wenhong Tian. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Large Language Models (LLMs) have demonstrated exceptional performance on natural language processing tasks. However, these models are computationally intensive and require substantial hardware resources for deployment. Quantization has emerged as a popular technique for LLM deployment, reducing memory requirements, but it results in accuracy degradation, particularly when using low-bit quantization. To mitigate this accuracy loss, we introduce Low-Rank Compensation (LoRC), a novel compensation mechanism that aims to recover the performance drop caused by quantization. Additionally, we propose Low-Rank Quantization (LoRQ), which further reduces the quantization-induced loss by adaptively adjusting weights at the element-wise level to help LLMs accommodate quantized computations. LoRC focuses on compensating for accuracy loss during inference, LoRQ integrates low-rank compensation directly into the quantization process, and they do not need end-to-end fine-tuning with LLM. Furthermore, we propose the Rank-α Addition Strategy (RαAS) to combine LoRC into the inference framework, which improves inference accuracy without increasing inference latency. Experimental results show that our method outperforms the state-of-the-art OmniQuant by 1.89% on several common zero-shot datasets under the W4A4 setting of the widely-used LLaMA. Through the joint design of algorithms and systems, our techniques can be easily integrated into the FlexGen inference framework without introducing additional inference latency, thereby maintaining high throughput while improving accuracy.

Current parameter-efficient fine-tuning methods for adapting pre-trained language models to downstream tasks are susceptible to interference from noisy data. Conventional noise-handling approaches either rely on laborious data pre-processing or employ model architecture modifications prone to error accumulation. In contrast to existing noise-process paradigms, we propose a noise-robust adaptation method via asymmetric LoRA poisoning experts (LoPE), a novel framework that enhances model robustness to noise only with generated noisy data. Drawing inspiration from the mixture-of-experts architecture, LoPE strategically integrates a dedicated poisoning expert in an asymmetric LoRA configuration. Through a two-stage paradigm, LoPE performs noise injection on the poisoning expert during fine-tuning to enhance its noise discrimination and processing ability. During inference, we selectively mask the dedicated poisoning expert to leverage purified knowledge acquired by normal experts for noise-robust output. Extensive experiments demonstrate that LoPE achieves strong performance and robustness purely through the low-cost noise injection, which completely eliminates the requirement of data cleaning.

This article proposes a cloud-edge–terminal collaborative resource allocation architecture that efficiently allocates resources. Traditional resource allocation often focuses solely on optimizing delay and service cost, making it less suitable for intensive real-world scenarios. In response, a comprehensive multiobjective resource allocation model is developed, encompassing delay, service cost, load balancing, and resource utilization. This article proposes a diversity-filling large-scale multiobjective evolutionary algorithm based on generative adversarial networks (DFGAN-LSMOEA). The Otsu-based grouping method in DFGAN-LSMOEA is employed to group decision variables and improve optimization performance. Compared with state-of-the-art algorithms, the proposed method validates its effectiveness and advantages in the applications, particularly when handling high-dimensional decision variables and dynamic demands.

Deep neural networks can provide environment sensing and decision support for vehicles in 6G intelligent autonomous transportation systems; however, the high computational cost associated with complex tasks limits the deployment of models on edge devices. To address this issue, this paper introduces a large-scale multiobjective filter pruning method, which stratifies the population by Angle Penalty Distance (APD) and selects the individuals to be updated. Meanwhile, the global search capability of the algorithm is enhanced by combining sampling update strategies and quantum behavioral update, in different situations. Moreover, a dynamic optimization tuning strategy is proposed to regulate the balance between exploration and exploitation within the algorithm. In the depth estimation task, the model is pruned using three objectives: root mean square error (RMSE), number of parameters, and FLOPs. Experimental results indicate that the pruned model obtains the minimum RMSE and the maximum compression ratio of the number of parameters and FLOPs, and can be effectively deployed in sensing-computing integrated chip and system for intelligent transportation systems.