Pumped storage power stations not only serve as a special load but also store excess electricity from the system, significantly reducing curtailment of wind and solar power. This dual function ensures stable operation grid enhances its economic benefits. The scheduling optimization problem combined wind–solar–pumped system is addressed in this study, an model proposed with objective maximizing total revenue. designed to comprehensively account for generation revenues power, photovoltaic...

Urban water demand forecasting is beneficial for reducing the waste of resources and enhancing environmental protection in sustainable management. However, it a challenging task to accurately predict affected by range factors with nonlinear uncertainty temporal patterns. This paper proposes new hybrid framework urban daily multiple variables, called attention-based CNN-LSTM model, which combines convolutional neural network (CNN), long short-term memory (LSTM), attention mechanism (AM),...

Abstract The solar eclipse on 15 January 2010 traversed Asia and completed its travel the Shandong Peninsula in China at sunset. Two vertical incidence ionosondes Wuhan Beijing oblique ionosonde network North were implemented to record ionospheric response eclipse. Following initial electron density decrease caused by eclipse, ionosphere was characterized a strong premidnight enhancement, subsequent decay, ~10 h later postmidnight enhancement. Neither geomagnetic disturbance occurred during...

It has recently been argued that AI models' representations are becoming aligned as their scale and performance increase. Empirical analyses have designed to support this idea conjecture the possible alignment of different toward a shared statistical model reality. In paper, we propose learning-theoretic perspective representation alignment. First, review connect notions based on metric, probabilistic, spectral ideas. Then, focus stitching, particular approach understanding interplay between...

Deep learning based on deep convolutional neural networks (CNNs) is extremely efficient in solving classification problems speech recognition, computer vision, and many other fields. But there no enough theoretical understanding about this topic, especially the generalization ability of induced CNN algorithms. In article, we develop some analysis a algorithm for binary with data spheres. An essential property problem lack continuity or high smoothness target function associated convex loss...

The Wuhan Ionospheric Oblique Backscattering Sounding System (WIOBSS) is a monostatic high-frequency sky-wave radar used for ionospheric remote sensing. sweep frequency backscatter ionogram (SFBI) recorded by the WIOBSS contains both echo scattered distant terrestrial surface and vertical incidence (VI) reflected local ionosphere over sounding station. approach SFBI inversion introduced in this letter requires input of leading edge peak height derived from VI echo. final output 2-D electron...

Nowadays, it becomes very convenient to collect synchronized WiFi received signal strength and inertial measurement (RSS+IMU) sequences by mobile devices, which enables the promising solution conduct unsupervised indoor localization without pain of radio-map calibration. To relax needs floor-map information or trajectory knowledge, this paper proposes learn a transitional model (TM), segments massive unlabeled train that captures expected relationship between {zt--1, zt } ut--1, where zt--1,...

Semantic parsing maps natural language (NL) utterances into logical forms (LFs), which underpins many advanced NLP problems. parsers gain performance boosts with deep neural networks, but inherit vulnerabilities against adversarial examples. In this paper, we provide the first empirical study on robustness of semantic in presence attacks. Formally, adversaries are considered to be perturbed utterance-LF pairs, whose have exactly same meanings as original ones. A scalable methodology is...

Aircraft detection is an important application of Wuhan Ionosonde Sounding System (WISS), which recently has been developed by the Ionospheric Laboratory University. Since ionosphere varies temporally and spatially, severe multipath effects are produced, jeopardise characteristic quantities extracting targets from recorded data. To solve above problems further identify fuzzy signals, this study presents a neural networks time‐frequency‐based algorithm. By networks, extracted data, then,...

Automatic text summarization is a key task in natural language processing. High-quality datasets can effectively promote the research progress of summarization. Recent closer to generate abstractive summarizations by using deep learning methods. However, there lack high-quality and large-scale available public. Besides, it difficult construct this kind dataset manually. The Tibetan still its infancy due public datasets. In order development informatization. we artificially constructed small...

Pairwise learning is widely employed in ranking, similarity and metric learning, area under the ROC curve (AUC) maximization, many other tasks involving sample pairs. with deep neural networks was considered for but enough theoretical understanding about this topic lacking. In letter, we apply symmetric to pairwise ranking a hinge loss ϕh carry out generalization analysis algorithm. A key step our characterize function that minimizes risk. This motivates us first find minimizer of ϕh-risk...

Recently, text-to-3D generation has attracted significant attention, resulting in notable performance enhancements. Previous methods utilize end-to-end 3D models to initialize Gaussians, multi-view diffusion enforce consistency, and text-to-image refine details with score distillation algorithms. However, these exhibit two limitations. Firstly, they encounter conflicts directions since different aim produce diverse assets. Secondly, the issue of over-saturation not been thoroughly...

Adaptive modulation communication systems have been popular nowadays. The tradeoff between the symbol error rate and data resulting from constellation option is crucial in adaptive modulation. In this paper, we propose a subset selection (CSS) approach to seek design novel efficient approximation algorithms tackle CSS problems. new theorems studies on algorithmic systematic aspects for are facilitated. Our attempt cope with problems would be valuable future adjustable sets.

科学家借助功能性磁共振成像技术(functional magnetic resonance imaging, fMRI) 将大脑划分成不同的体素, 并通过不同体素的应答强弱 来分析不同的脑活动模式.通过影像学方法对大脑进行解码, 人们能够感知和探索所处的视觉环境, 进而为理解和破译大脑的运作模式带来了便利.现有的脑影像智能分析方法主要采用的机器学习和数据挖掘算法包括分类 [5] 、回归 [6] 、关联规则 挖掘 [7] 、特征分析 [8] 等. 在大数据背景下, 基于智能方法的脑影像分析技术将面临新的挑战, 具体有 以下几个方面: 不同模态的数据 (包括神经影像数据、基因数据和生理行为数据等) 往往是异构的, 如 何综合互补地利用每个模态的信息是一个急需解决的关键问题; 疾病相关数据维度极高、结构复杂、 具有小样本特性, 如何高效地对这些大规模复杂数据进行有效分析, 从大量特征中找到与脑疾病相关 的特征集合作为生物标志物来指导临床诊断是另一个难题; 不同的人脑具有自身独有的脑活动模式, 如何设计一种高效的配准方法, 将不同个体的功能性磁共振成像技术 (fMRI)...