Data missing remains a difficult and important problem in the transportation information system, which seriously restricts application of intelligent system (ITS), dominatingly on traffic monitoring, e.g., data collection, state estimation, control. Numerous imputation methods had been proposed last decade. However, lacking sufficient temporal variation characteristic analysis as well spatial correlation measurements leads to limited completion precision, poses major challenge for an ITS....

Machine learning (ML) technology has shown its unique advantages in many fields and excellent performance applications, such as image recognition, speech recommendation systems, natural language processing. Recently, the applicability of ML wireless sensor networks (WSNs) attracted much attention. As resources are limited WSNs, identifying how to improve resource utilization achieve power-efficient load balancing is becoming a critical issue WSNs. Traditional green routing algorithms aim...

Millions of smart home speakers, such as Amazon Echo and Google Home, have been purchased by U.S. consumers. However, the security privacy speakers not rigorously examined, which raise critical concerns. In this paper, we investigate untold severe leakage speakers. Specifically, examine a new passive attack, referred to voice command fingerprinting on We demonstrate that attacker, who can only eavesdrop encrypted traffic between speaker cloud server, infer users' commands compromise millions...

With the access of massive mobile devices, spectrum resources are becoming increasingly scarce. How to effectively and securely utilize limited has become a fundamental challenge for future communication systems. Focusing on intelligent sensing sharing, this article proposes blockchain-based two-stage secure sharing auction mechanism (BISA), which selects appropriate base stations form consortium blockchain guarantee efficient with low complexity. In first stage, reverse-auction-based...

Characterizing regulatory effects of genomic variants in plants remains a challenge. Although several tools based on deep-learning models and large-scale chromatin-profiling data have been available to predict elements variant effects, no dedicated or web services reported plants. Here, we present PlantDeepSEA as deep learning-based service multiple tissues six plant species (including four crops). provides two main functions. One is called Variant Effector, which aims the sequence chromatin...

Federated learning (FL) provides a promising solution to meet the requirements of data privacy and security in intelligent transportation systems (ITS), which enables edge devices road side units (RSUs) collaboratively train models without exposing raw data. However, deep leakage from gradients (DLG) still leads risk divulging original Meanwhile, existing gradient protection methods based on secure multi-party computation (SMC) result huge communication overheads latency, are difficult...

Rebar constitutes a crucial element within tunnel lining structures, where its precise arrangement plays pivotal role in determining both structural stability and load-bearing capacity. Due to the rebar’s high dielectric constant approaching infinity, radar signal reflections are intensified, manifesting as distinct hyperbolic patterns imagery. By performing convolutional operations, these features of rebar can be effectively extracted from images. Building upon feature extraction...

Invariant molecular representation models provide potential solutions to guarantee accurate prediction of properties under distribution shifts out-of-distribution (OOD) by identifying and leveraging invariant substructures inherent the molecules. However, due complex entanglement functional groups frequent display activity cliffs properties, separation molecules becomes inaccurate tricky. This results in inconsistent semantics among identified existing models, which means sharing identical...

Accurately determining the geographic location of an Internet host is important for location-aware applications such as location-based advertising and network diagnostics. Despite their fast response time, widely used database-driven geolocation approaches provide only inaccurate locations. Delay measurement based improve estimation accuracy but still suffer from a limited precision (about 10 km) long time (tens seconds) to localize single PC, which cannot meet demand precise real-time...

Abstract As a signaling molecule, nitric oxide (NO) regulates the development and stress response in different organisms. The major biological activity of NO is protein S‐nitrosylation, whose function fungi remains largely unclear. Here, it found rice blast fungus Magnaporthe oryzae , de‐nitrosylation process essential for functional appressorium formation during infection. Nitrosative caused by excessive accumulation harmful fungal While S‐nitrosoglutathione reductase GSNOR‐mediated removes...

Transformers have emerged as the architecture of choice for many state-of-the-art AI models, showcasing exceptional performance across a wide range applications. However, memory demands imposed by limit their ability to handle long sequences, thereby posing challenges in utilizing videos, actions, and other long-form sequences modalities complex environments. We present novel approach, Ring Attention with Blockwise (Ring Attention), which leverages blockwise computation self-attention...

We introduce a new unsupervised pretraining objective for reinforcement learning. During the reward-free phase, agent maximizes mutual information between tasks and states induced by policy. Our key contribution is novel lower bound of this intractable quantity. show that reinterpreting combining variational successor features~\citep{Hansen2020Fast} with nonparametric entropy maximization~\citep{liu2021behavior}, can be efficiently optimized. The proposed method Active Pretraining Successor...

Due to the distributed and dynamic characteristics of Internet Vehicles (IoV) continuous growth in number devices, content-centric decentralized vehicular named data networking (VNDN) has become more suitable for content-oriented applications IoV. However, existing centralized architecture is prone failure single points, which results trust problems key verification between cross-domain nodes consuming power reducing lifetime. Focusing on secure management power-efficient routing, this...

Vertical federated learning (VFL) is a promising category of for the scenario where data vertically partitioned and distributed among parties. VFL enriches description samples using features from different parties to improve model capacity. Compared with horizontal learning, in most cases, applied commercial cooperation companies. Therefore, contains tremendous business values. In past few years, has attracted more attention both academia industry. this paper, we systematically investigate...

In order to reduce traffic conflicts on cross-intensive roads, this paper proposes a new early warning system based the active safety concept. The collects real-time vehicle data using roadside sensors and transmits results drivers major road in timely manner via lights. research, principles of are discussed detail, including how dynamics collected potential collisions identified avoided. Through driving simulation experiment, speed prediction model after implementation was examined. Results...

Artificial neural networks are widely used in various fields, such as intelligent road networks, Internet of Things, and smart medical systems due to their ability process large amounts data parallel, store information a distributed manner, self-organize self-learn. Cloud computing technology has further expanded the development network applications. However, user often contains sensitive information, once management right is transferred cloud, it faces serious security privacy issues. In...

Cognitive radio (CR) provides an effective solution to meet the huge bandwidth requirements in intelligent transportation systems (ITS), which enables secondary users (SUs) access idle spectrum of primary (PUs). However, high mobility and real-time service result additional transmission collisions interference, degrades rate quality (QoS) ITS. This paper proposes a algorithm (Feilin) based on federated deep reinforcement learning (FDRL) improve rate, maximizes QoS reward function with...

This Nowadays, people are paying more and attention to the quality of home environment with rapid popularity smart system Internet things technology. In this paper, a new method based on technology WSN is adopted monitor some indicators indoors, so that damaging caused by these imperceptible even dangerous indications, such as PM 2.5, temperature, humidity concentration carbon monoxide, can be reduced or eliminated. The proposed 2.5 detector used in acquiring real-time data for surveillance...

For the traditional inspection methods, visual data is firstly recorded on forms and then input manually into computer, which inefficient creates errors frequently. This research aims at establishing a smartphone-based bridge management system that can avoid such inputting facilitate process. The enables inspector to complete information collection in portable smart phone. site photos related investigated structures be easily added edited during work with help of After investigation, report...

The development of an intelligent transportation system (ITS) architecture has attracted increasing attention across the world because such provides a common reference for ITS community. Standardization procedures in is necessary and urgent architecture's inherent complexity, which requires involvement several teams. Although computer-aided systems engineering model widely adopted to accomplish parts task, so far no countries define strict, clear, applicable procedure aid This paper first...

Website fingerprinting can reveal which sensitive website a user visits over encrypted network traffic. Obfuscating traffic, e.g., adding dummy packets, is considered as primary approach to defend against fingerprinting. How-ever, existing defenses relying on traffic obfuscation are either ineffective or introduce significant overheads. As recent attacks heavily rely deep neural networks achieve high accuracy, producing adversarial examples could be utilized new way obfuscate Unfortunately,...