MotionQ: Operator-Conditioned Motion Quotients for Cross-Observation WiFi Gesture Recognition
为解决WiFi手势识别在不同布局和观测配置下的准确性问题,提出MotionQ方法,通过生成操作条件下的运动测量并保留关键信息来提高鲁棒性。
为解决WiFi手势识别在不同布局和观测配置下的准确性问题,提出MotionQ方法,通过生成操作条件下的运动测量并保留关键信息来提高鲁棒性。
本文提出BReF方法,通过比较语言模型在文本扰动下的概率分布变化来识别模型的来源,解决了细粒度溯源难题。
This work addresses the challenge that lightweight models often struggle to balance parameter efficiency and predictive performance in long-term time series forecasting, primarily due to non-stationarity, high-frequency perturbations, and cross-period dependencies. Building upon the SparseTSF framework, the paper introduces three key innovations: trend-aware reversible instance normalization to mitigate distributional shifts, scale-adaptive gated denoising to suppress high-frequency noise, and a multi-scale gated attention MLP to enhance modeling of cross-period features. Extensive experiments demonstrate that the proposed method achieves consistently superior forecasting accuracy across multiple benchmarks. Ablation studies further confirm that each component effectively improves distribution adaptability, input robustness, and cross-period representational capacity, respectively.
This work addresses the longstanding trade-off between computational efficiency and modeling capacity in time series anomaly detection. The authors propose a lightweight model based on chunked representation learning that jointly captures hierarchical temporal patterns and complex inter-variable dependencies through a multi-receptive-field convolutional backbone, multi-scale adaptive attention aggregation, and an explicit cross-variable fusion mechanism. To enhance feature discriminability and generalization, the model incorporates a temporal chunk ordering pretext task and triplet loss during pretraining. Evaluated on the TSB-AD benchmark, the method achieves state-of-the-art accuracy in both univariate and multivariate settings—significantly outperforming PaAno in metrics such as VUS-PR—while maintaining low computational overhead, making it well-suited for real-time inference in resource-constrained environments.
This study addresses the lack of a general characterization for BCH codes and LCD cyclic codes of length $n = \lambda(q^m + 1)$, where $\lambda \mid q - 1$. By fully describing the structure of $q$-cyclotomic cosets modulo $n$, the work extends existing results—previously limited to the case $\lambda = 1$—to arbitrary $\lambda$ dividing $q - 1$. Leveraging finite field theory, cyclotomic coset analysis, and algebraic coding techniques, the paper explicitly identifies the largest coset representatives for odd $m$ and establishes necessary and sufficient conditions for a BCH code to be self-dual. Key contributions include precise dimension formulas for several families of BCH codes, an improved lower bound on the minimum distance of certain BCH codes up to $2(\delta+1)$—partially achieving optimality—and an exact enumeration of all LCD cyclic codes of the specified length.
为解决WiFi手势识别在不同布局和观测配置下的准确性问题,提出MotionQ方法,通过生成操作条件下的运动测量并保留关键信息来提高鲁棒性。
本文提出BReF方法,通过比较语言模型在文本扰动下的概率分布变化来识别模型的来源,解决了细粒度溯源难题。
This work addresses the challenge that lightweight models often struggle to balance parameter efficiency and predictive performance in long-term time series forecasting, primarily due to non-stationarity, high-frequency perturbations, and cross-period dependencies. Building upon the SparseTSF framework, the paper introduces three key innovations: trend-aware reversible instance normalization to mitigate distributional shifts, scale-adaptive gated denoising to suppress high-frequency noise, and a multi-scale gated attention MLP to enhance modeling of cross-period features. Extensive experiments demonstrate that the proposed method achieves consistently superior forecasting accuracy across multiple benchmarks. Ablation studies further confirm that each component effectively improves distribution adaptability, input robustness, and cross-period representational capacity, respectively.
This work addresses the longstanding trade-off between computational efficiency and modeling capacity in time series anomaly detection. The authors propose a lightweight model based on chunked representation learning that jointly captures hierarchical temporal patterns and complex inter-variable dependencies through a multi-receptive-field convolutional backbone, multi-scale adaptive attention aggregation, and an explicit cross-variable fusion mechanism. To enhance feature discriminability and generalization, the model incorporates a temporal chunk ordering pretext task and triplet loss during pretraining. Evaluated on the TSB-AD benchmark, the method achieves state-of-the-art accuracy in both univariate and multivariate settings—significantly outperforming PaAno in metrics such as VUS-PR—while maintaining low computational overhead, making it well-suited for real-time inference in resource-constrained environments.
This study addresses the lack of a general characterization for BCH codes and LCD cyclic codes of length $n = \lambda(q^m + 1)$, where $\lambda \mid q - 1$. By fully describing the structure of $q$-cyclotomic cosets modulo $n$, the work extends existing results—previously limited to the case $\lambda = 1$—to arbitrary $\lambda$ dividing $q - 1$. Leveraging finite field theory, cyclotomic coset analysis, and algebraic coding techniques, the paper explicitly identifies the largest coset representatives for odd $m$ and establishes necessary and sufficient conditions for a BCH code to be self-dual. Key contributions include precise dimension formulas for several families of BCH codes, an improved lower bound on the minimum distance of certain BCH codes up to $2(\delta+1)$—partially achieving optimality—and an exact enumeration of all LCD cyclic codes of the specified length.