Zekun Li, Jifan Yu, Haoxuan Li, Ye He, Daniel Zhang-Li, Shangqing Tu, Joy Jia Yin Lim, Yikun Jiang, Jiaxin Yuan, Yu Zhang. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Score normalization is a key back-end technique in speaker verification for improving score comparability across trials. As a convenient and widely used normalization method, Adaptive Symmetric Normalization (AS-Norm) standardizes raw scores using mean-and-standard-deviation normalization parameters estimated from an adaptive cohort comprising the most similar impostor scores. However, since the adaptive cohort retains only the top-ranked impostor scores, these central-moment statistics may not optimally characterize the empirical distribution of these scores, resulting in suboptimal speaker verification performance. In this letter, we propose Adaptive Symmetric Extreme Value Normalization (AS-EVNorm), which treats the adaptive cohort as upper-tail samples from the impostor-score distribution and models them under extreme-value theory for more accurate normalization parameters. Experiments on VoxCeleb and CN-Celeb show that AS-EVNorm consistently reduces both EER and minDCF compared with AS-Norm across a broad range of adaptive cohort configurations, while maintaining competitive normalization time.

The deep neural networks combined with the statistical shape model have enabled efficient deformable 2D-3D registration and recovery of 3D anatomical structures from a single radiograph. However, the recovered volumetric image tends to lack the volumetric fidelity of fine-grained anatomical structures and explicit consideration of cross-dimensional semantic correspondence. In this paper, we introduce a simple but effective solution for semantically-consistent deformable 2D-3D registration and detailed volumetric image recovery by inferring a voxel-wise registration field between the cone-beam computed tomography and a single lateral cephalometric radiograph (LC). The key idea is to refine the initial statistical model-based registration field with craniofacial structural details and semantic consistency from the LC. Specifically, our framework employs a self-supervised scheme to learn a voxel-level refiner of registration fields to provide fine-grained craniofacial structural details and volumetric fidelity. We also present a weakly supervised semantic consistency measure for semantic correspondence, relieving the requirements of volumetric image collections and annotations. Experiments showcase that our method achieves deformable 2D-3D registration with performance gains over state-of-the-art registration and radiograph-based volumetric reconstruction methods. The source code is available at https://github.com/Jyk-122/SC-DREG.

Large language models have shown an impressive societal impact owing to their excellent understanding and logical reasoning skills. However, such strong ability relies on a huge amount of computing resources, which makes it difficult to deploy LLMs on computing resource-constrained platforms. Currently, LLMs process each token equivalently, but we argue that not every word is equally important. Some words should not be allocated excessive computing resources, particularly for dispensable terms in simple questions. In this paper, we propose a novel dynamic inference paradigm for LLMs, namely D-LLMs, which adaptively allocate computing resources in token processing. We design a dynamic decision module for each transformer layer that decides whether a network unit should be executed or skipped. Moreover, we tackle the issue of adapting D-LLMs to real-world applications, specifically concerning the missing KV-cache when layers are skipped. To overcome this, we propose a simple yet effective eviction policy to exclude the skipped layers from subsequent attention calculations. The eviction policy not only enables D-LLMs to be compatible with prevalent applications but also reduces considerable storage resources. Experimentally, D-LLMs show superior performance, in terms of computational cost and KV storage utilization. It can reduce up to 45\% computational cost and KV storage on Q\&A, summarization, and math solving tasks, 50\% on commonsense reasoning tasks.

With the development of Internet of Things(IoT), more and more smart devices are connected into the Internet. The security and privacy issues of IoT devices have received increasingly academic and industrial attentions. Vulnerability detection is the key technology to protect IoT devices from zero-day attacks. However, traditional methods and tools of vulnerability detection cannot be directly used in analyzing IoT firmware. This paper firstly reviews related works on vulnerability detection in IoT firmware, previous researches are classified into four types i.e. static analysis, symbolic execution, fuzzing on emulators and comprehensive testing. Then, this paper points out that the specificity of vulnerability detection in IoT firmware is to detect logical flaws in embedded binaries which are built on the MIPS architecture. Finally, this paper proposes a method based on fuzzing and static analysis to detect authentication bypass flaws in IoT embedded binary servers. The proposed method is proved to be effective by verifying known CVEs as well as discovering unknown ones.