PunGraph: Retrieval-Enhanced Phonetic-Semantic Graph Reasoning for Pun Understanding
本文提出PunGraph,一种结合检索增强的语音-语义图推理框架,用于解决双关语理解问题,通过构建语音-语义词汇图并利用WebPun数据集来提高小规模语言模型的解释能力。
本文提出PunGraph,一种结合检索增强的语音-语义图推理框架,用于解决双关语理解问题,通过构建语音-语义词汇图并利用WebPun数据集来提高小规模语言模型的解释能力。
针对多任务学习中目标函数尺度不同导致的优化偏差问题,提出了一种基于对数变换的尺度不变价值函数标量化方法SIMS。
本文提出广义代理迭代框架,统一描述迭代策略改进与递归自我改进,通过调整机制和标准来区分不同实例,为分析设计新系统提供基础。
This work addresses the limitations of existing RGB-IR object detection methods for unmanned aerial vehicles, which typically rely on static fusion strategies that fail to account for spatially varying modality reliability. To overcome this, the authors propose EGM-Det, a dual-stream framework that preserves modality-specific representations and introduces an entropy-guided offset gating mechanism. This mechanism leverages input intensity, local entropy, and cross-modal discrepancies to construct shallow entropy priors, dynamically guiding multi-scale spatial-channel alignment and fusion. Additionally, an entropy-adaptive supervised cross-modal knowledge distillation strategy is designed to optimize training. The proposed method achieves state-of-the-art performance across three benchmarks—DroneVehicle, LLVIP, and VEDAI—with a notable improvement of over 10 percentage points in mAP on VEDAI.
This study addresses a critical limitation of existing integer-valued matrix autoregressive models, which cannot accommodate negative integers and thus fail to model real-world scenarios involving differenced series or financial tick data. To overcome this, the authors propose the Z-MINAR model—the first matrix autoregressive framework defined over the full set of integers, both positive and negative. The approach introduces a signed matrix sparsity operator and innovation terms based on an extended Poisson distribution, thereby preserving the underlying matrix topology. Parameter estimation is achieved via projected conditional least squares. Theoretical analysis establishes the model’s stationarity, causality, and asymptotic normality. Simulations demonstrate that Z-MINAR substantially outperforms existing methods in estimation accuracy, robustness, and adaptability, while empirical application successfully uncovers the dynamic spatiotemporal dependence structure in urban crime count data.
本文提出PunGraph,一种结合检索增强的语音-语义图推理框架,用于解决双关语理解问题,通过构建语音-语义词汇图并利用WebPun数据集来提高小规模语言模型的解释能力。
针对多任务学习中目标函数尺度不同导致的优化偏差问题,提出了一种基于对数变换的尺度不变价值函数标量化方法SIMS。
本文提出广义代理迭代框架,统一描述迭代策略改进与递归自我改进,通过调整机制和标准来区分不同实例,为分析设计新系统提供基础。
This work addresses the limitations of existing RGB-IR object detection methods for unmanned aerial vehicles, which typically rely on static fusion strategies that fail to account for spatially varying modality reliability. To overcome this, the authors propose EGM-Det, a dual-stream framework that preserves modality-specific representations and introduces an entropy-guided offset gating mechanism. This mechanism leverages input intensity, local entropy, and cross-modal discrepancies to construct shallow entropy priors, dynamically guiding multi-scale spatial-channel alignment and fusion. Additionally, an entropy-adaptive supervised cross-modal knowledge distillation strategy is designed to optimize training. The proposed method achieves state-of-the-art performance across three benchmarks—DroneVehicle, LLVIP, and VEDAI—with a notable improvement of over 10 percentage points in mAP on VEDAI.
This study addresses a critical limitation of existing integer-valued matrix autoregressive models, which cannot accommodate negative integers and thus fail to model real-world scenarios involving differenced series or financial tick data. To overcome this, the authors propose the Z-MINAR model—the first matrix autoregressive framework defined over the full set of integers, both positive and negative. The approach introduces a signed matrix sparsity operator and innovation terms based on an extended Poisson distribution, thereby preserving the underlying matrix topology. Parameter estimation is achieved via projected conditional least squares. Theoretical analysis establishes the model’s stationarity, causality, and asymptotic normality. Simulations demonstrate that Z-MINAR substantially outperforms existing methods in estimation accuracy, robustness, and adaptability, while empirical application successfully uncovers the dynamic spatiotemporal dependence structure in urban crime count data.