623509 research outputs found
Sort by
Analysis of ELSA COVID-19 Substudy response rate using machine learning algorithms
No full textNational Statistical Organisations every year spend time and money to collect information through surveys. Some of these surveys include follow-up studies, and usually, some participants due to factors such as death, immigration, change of employment, health, etc, do not participate in future surveys. In this study, we focus on the English Longitudinal Study of Ageing (ELSA) COVID-19 Substudy, which was carried out during the COVID-19 pandemic in two waves. In this substudy, some participants from wave 1 did not participate in wave 2. Our purpose is to predict non-responses using Machine Learning (ML) algorithms such as K-nearest neighbours (KNN), random forest (RF), AdaBoost, logistic regression, neural networks (NN), and support vector classifier (SVC). We find that RF outperforms other models in terms of balanced accuracy, KNN in terms of precision and test accuracy, and logistics regressions in terms of the area under the receiver operating characteristic curve (ROC), i.e. AUC
Perfect Matchings and Popularity in the Many-to-Many Setting
No full textWe consider a matching problem in a bipartite graph where every vertex has a capacity and a strict preference order on its neighbors. Furthermore, there is a cost function on the edge set. We assume admits a perfect matching, i.e., one that fully matches all vertices. It is only perfect matchings that are feasible for us and we are interested in those perfect matchings that are popular within the set of perfect matchings. It is known that such matchings (called popular perfect matchings) always exist and can be efficiently computed. What we seek here is not any popular perfect matching, but a min-cost one. We show a polynomial-time algorithm for finding such a matching; this is via a characterization of popular perfect matchings in in terms of stable matchings in a colorful auxiliary instance. This is a generalization of such a characterization that was known in the one-to-one setting
Improving Viewpoint-Independent Object-Centric Representations through Active Viewpoint Selection
No full textGiven the complexities inherent in visual scenes, such as object occlusion, a comprehensive understanding often requires observation from multiple viewpoints. Existing multi-viewpoint object-centric learning methods typically employ random or sequential viewpoint selection strategies. While applicable across various scenes, these strategies may not always be ideal, as certain scenes could benefit more from specific viewpoints. To address this limitation, we propose a novel active viewpoint selection strategy. This strategy predicts images from unknown viewpoints based on information from observation images for each scene. It then compares the object-centric representations extracted from both viewpoints and selects the unknown viewpoint with the largest disparity, indicating the greatest gain in information, as the next observation viewpoint. Through experiments on various datasets, we demonstrate the effectiveness of our active viewpoint selection strategy, significantly enhancing segmentation and reconstruction performance compared to random viewpoint selection. Moreover, our method can accurately predict images from unknown viewpoints
Inference-to-complete: A High-performance and Programmable Data-plane Co-processor for Neural-network-driven Traffic Analysis
No full textNeural-networks-driven intelligent data-plane (NN-driven IDP) is becoming an emerging topic for excellent accuracy and high performance. Meanwhile we argue that NN-driven IDP should satisfy three design goals: the flexibility to support various NNs models, the low-latency-high-throughput inference performance, and the data-plane-unawareness harming no performance and functionality. Unfortunately, existing work either over-modify NNs for IDP, or insert inline pipelined accelerators into the data-plane, failing to meet the flexibility and unawareness goals.
In this paper, we propose Kaleidoscope, a flexible and high-performance co-processor located at the bypass of the data-plane. To address the challenge of meeting three design goals, three key techniques are presented. The programmable run-to-completion accelerators are developed for flexible inference. To further improve performance, we design a scalable inference engine which completes low-latency and low-cost inference for the mouse flows, and perform complex NNs with high-accuracy for the elephant flows. Finally, raw-bytes-based NNs are introduced, which help to achieve unawareness. We prototype Kaleidoscope on both FPGA and ASIC library. In evaluation on six NNs models, Kaleidoscope reaches 256-352 ns inference latency and 100 Gbps throughput with negligible influence on the data-plane. The on-board tested NNs perform state-of-the-art accuracy among other NN-driven IDP, exhibiting the the significant impact of flexibility on enhancing traffic analysis accuracy.Under revie
Cityscape-Adverse: Benchmarking Robustness of Semantic Segmentation with Realistic Scene Modifications via Diffusion-Based Image Editing
No full textRecent advancements in generative AI, particularly diffusion-based image editing, have enabled the transformation of images into highly realistic scenes using only text instructions. This technology offers significant potential for generating diverse synthetic datasets to evaluate model robustness. In this paper, we introduce Cityscape-Adverse, a benchmark that employs diffusion-based image editing to simulate eight adverse conditions, including variations in weather, lighting, and seasons, while preserving the original semantic labels. We evaluate the reliability of diffusion-based models in generating realistic scene modifications and assess the performance of state-of-the-art CNN and Transformer-based semantic segmentation models under these challenging conditions. Additionally, we analyze which modifications have the greatest impact on model performance and explore how training on synthetic datasets can improve robustness in real-world adverse scenarios. Our results demonstrate that all tested models, particularly CNN-based architectures, experienced significant performance degradation under extreme conditions, while Transformer-based models exhibited greater resilience. We verify that models trained on Cityscape-Adverse show significantly enhanced resilience when applied to unseen domains. Code and datasets will be released at https://github.com/naufalso/cityscape-adverse.19 pages, under review, code and dataset will be available at https://github.com/naufalso/cityscape-advers
New bounds for fundamental Fourier coefficients of Siegel modular forms
No full textWe prove new bounds for the Fourier coefficients of Jacobi forms using a method of Iwaniec. In view of the Fourier-Jacobi expansion of degree two Siegel modular forms, we can use these to obtain strong bounds on fundamental Fourier coefficients of Siegel modular forms.23 page
Local well-posedness for a fourth-order nonlinear dispersive system on the 1D torus
No full textThis paper is concerned with the initial value problem for a system of one-dimensional fourth-order dispersive partial differential equations on the torus with nonlinearity involving derivatives up to second order. This paper gives sufficient conditions on the coefficients of the system for the initial value problem to be time-locally well-posed in Sobolev spaces with high regularity. The proof is based on the energy method combined with the idea of a gauge transformation and the technique of Bona-Smith type parabolic regularization. The sufficient conditions can been found in connection with geometric analysis on a fourth-order geometric dispersive partial differential equation for curve flows on a compact locally Hermitian symmetric space.37 page
Contramodules for algebraic groups: induction and projective covers
No full textIn this paper we will investigate contramodules for algebraic groups. Namely, we give contra-analogs to two 20th century results about comodules. Firstly, we show that induction of contramodules over coordinate rings of algebraic groups is exact if and only if the associated quotient variety is affine. Secondly, we give an inverse limit theorem for constructing projective covers of simple -modules using -structures of projective covers of simple modules for the first Frobenius kernel, .Updated to essentially match journal submission. Includes restructuring of results, typo fixes etc. 19 page
Continuum excitations in a spin-supersolid on a triangular lattice
No full textMagnetic, thermodynamic, neutron diffraction and inelastic neutron scattering are used to study spin correlations in the easy-axis XXZ triangular lattice magnet K2Co(SeO3)2. Despite the presence of quasi-2D supersolid magnetic order, the low-energy excitation spectrum contains no sharp modes and is instead a broad and structured multi-particle continuum. Applying a weak magnetic field drives the system into an m = 1/3 fractional magnetization plateau phase and restores sharp spin wave modes. To some extent, the behavior at zero field can be understood in terms of spin wave decay. However, the presence of clear excitation minima at the M-points of the Brillouin zone suggest that the spinon language may provide a more adequate description, and signals a possible proximity to a Dirac spin liquid state
Ergotropy and capacity optimization in Heisenberg spin-chain quantum batteries
No full textThis study examines the performance of finite spin quantum batteries (QBs) using Heisenberg spin models with Dzyaloshinsky-Moriya (DM) and Kaplan--Shekhtman--Entin-Wohlman--Aharony (KSEA) interactions. The QBs are modeled as interacting quantum spins in local inhomogeneous magnetic fields, inducing variable Zeeman splitting. We derive analytical expressions for the maximal extractable work, ergotropy and the capacity of QBs, as recently examined by Yang et al. [Phys. Rev. Lett. 131, 030402 (2023)]. These quantities are analytically linked through certain quantum correlations, as posited in the aforementioned study. Different Heisenberg spin chain models exhibit distinct behaviors under varying conditions, emphasizing the importance of model selection for optimizing QB performance. In antiferromagnetic (AFM) systems, maximum ergotropy occurs with a Zeeman splitting field applied to either spin, while ferromagnetic (FM) systems benefit from a uniform Zeeman field. Temperature significantly impacts QB performance, with ergotropy in the AFM case being generally more robust against temperature increases compared to the FM case. Incorporating DM and KSEA couplings can significantly enhance the capacity and ergotropy extraction of QBs. However, there exists a threshold beyond which additional increases in these interactions cause a sharp decline in capacity and ergotropy. This behavior is influenced by temperature and quantum coherence, which signal the occurrence of a sudden phase transition. The resource theory of quantum coherence proposed by Baumgratz et al. [Phys. Rev. Lett. 113, 140401 (2014)] plays a crucial role in enhancing ergotropy and capacity. However, ergotropy is limited by both the system\u27s capacity and the amount of coherence. These findings support the theoretical framework of spin-based QBs and may benefit future research on quantum energy storage devices.14 pages, 11 figure