We therefore pursue a data-driven alternative based on complex-valued V AEs and latent-space OOD scores. In recent years, data-driven approaches have emerged to alleviate the need for precise clutter modeling. Among them, V AEs [4] have demonstrated promising capabilities for anomaly and OOD detection in diverse applications, including radar detection [5], speech enhancement [6], medical imaging [7], industrial monitoring [8], and acoustic signal analysis [9]. These models learn a latent representation of the training data and use reconstruction or probabilistic criteria to detect deviations. Despite their effectiveness, most V AE-based detectors operate in the real domain and often treat complex-valued radar data by separating real and imaginary components into distinct channels. Recent advances in Complex-V alued Neural Networks (CVNNs) have shown the benefits of directly modeling complex-valued signals [10, 11].

Neural Information Processing Systems http://nips.cc/

Therefore, KLD has affine invariance. Part of the work was done during an internship at Huawei Inc. Correspondence author is Junchi Y an. Do the main claims made in the abstract and introduction accurately reflect the paper's Did you discuss any potential negative societal impacts of your work? If you ran experiments... (a) Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [Y es] (b) Did you specify all the training details (e.g., data splits, hyperparameters, how they Did you report error bars (e.g., with respect to the random seed after running experiments multiple times)? Did you include the total amount of compute and the type of resources used (e.g., type Did you include any new assets either in the supplemental material or as a URL? [No] Did you discuss whether and how consent was obtained from people whose data you're Did you discuss whether the data you are using/curating contains personally identifiable information or offensive content?

As a fundamental building block for visual analysis across aerial images, scene text etc., rotated object detection has recently been developed rapidly [

Speculative decoding accelerates large language model inference, but its reliance on a fixed speculation length is suboptimal in large-batch serving environments with diverse requests. This paper explores a new direction for dynamic adaptation by investigating a novel class of post-hoc, diagnostic signals. We propose Dynamic Speculative Decoding Engine (DSDE), a training-free framework built on two primary components: (1) a predictive signal based on the variance of the Kullback-Leibler (KLD) divergence, which diagnoses the generation's regional stability, and (2) an adaptive speculation length cap to mitigate the straggler problem in per-sequence decoding. Experiments demonstrate the potential of using KLD-based stability signals for dynamic adaptation. An algorithm guided by these signals achieves end-to-end latency competitive with leading baselines and exhibits superior robustness across diverse workloads. This robustness is particularly valuable in challenging low-acceptance-rate regimes, where the proposed signal maintains its diagnostic utility. Collectively, these findings validate post-hoc signals as a valuable component for building more robust and intelligent LLM inference systems, and highlight a promising direction for future research on dynamic speculation length adaptation.

We thank the reviewers for their thoughtful and constructive reviews. 'theoretically and experimentally grounded,' and'extremely well-written,' that it'could easily be used in practice,' and Below we respond to the major comments; we will fix the minor ones in the final version. Experiments for CIFAR-100 are in progress and we will add those and the results of Table I to the final paper. The MLP and CNN are a bit old models [...] We used MLP and CNNs since they were used in studies that we We will add the'mixeup training' results into the revised paper. We performed experiments similar to arXiv:1902.02476

The geometric Jensen--Shannon divergence (G-JSD) gained popularity in machine learning and information sciences thanks to its closed-form expression between Gaussian distributions. In this work, we introduce an alternative definition of the geometric Jensen--Shannon divergence tailored to positive densities which does not normalize geometric mixtures. This novel divergence is termed the extended G-JSD as it applies to the more general case of positive measures. We report explicitly the gap between the extended G-JSD and the G-JSD when considering probability densities, and show how to express the G-JSD and extended G-JSD using the Jeffreys divergence and the Bhattacharyya distance or Bhattacharyya coefficient. The extended G-JSD is proven to be a $f$-divergence which is a separable divergence satisfying information monotonicity and invariance in information geometry. We derive corresponding closed-form formula for the two types of G-JSDs when considering the case of multivariate Gaussian distributions often met in applications. We consider Monte Carlo stochastic estimations and approximations of the two types of G-JSD using the projective $γ$-divergences. Although the square root of the JSD yields a metric distance, we show that this is not anymore the case for the two types of G-JSD. Finally, we explain how these two types of geometric JSDs can be interpreted as regularizations of the ordinary JSD.

Therefore, KLD has affine invariance. Part of the work was done during an internship at Huawei Inc. Correspondence author is Junchi Y an. Do the main claims made in the abstract and introduction accurately reflect the paper's Did you discuss any potential negative societal impacts of your work? If you ran experiments... (a) Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [Y es] (b) Did you specify all the training details (e.g., data splits, hyperparameters, how they Did you report error bars (e.g., with respect to the random seed after running experiments multiple times)? Did you include the total amount of compute and the type of resources used (e.g., type Did you include any new assets either in the supplemental material or as a URL? [No] Did you discuss whether and how consent was obtained from people whose data you're Did you discuss whether the data you are using/curating contains personally identifiable information or offensive content?

As a fundamental building block for visual analysis across aerial images, scene text etc., rotated object detection has recently been developed rapidly [