Traffic Prediction With Missing Data: A Multi-Task Learning Approach

Wang, Ao; Ye, Yongchao; Song, Xiaozhuang; Zhang, Shiyao; Yu, James J.Q.

doi:10.1109/tits.2022.3233890

Cited by 15 publications

(5 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Additionally, data-driven approaches have gained widespread usage in the transportation sector. A study performed by the University of California, Berkeley, discovered that through the real-time analysis of traffic data, traffic flow can be optimized, congestion mitigated, and overall traffic efficiency enhanced (Zhang et al, 2017).…”

Section: The Achievements Of Data-driven Methodologies In Several Dom...mentioning

confidence: 99%

Building a sustainable peace and development model through data-driven Chinese peacekeeping actions based on UN Global Pulse

Yan,

Zhang

2024

J. Infras. Policy. Dev.

View full text Add to dashboard Cite

This paper aims to explore how to build a sustainable peace and development model for China’s peacekeeping efforts through the application of data-driven methods from UN Global Pulse. UN Global Pulse is a United Nations agency dedicated to using big data and artificial intelligence technologies to address global challenges. In this paper, we will introduce the working principles of UN Global Pulse and its application in the fields of peacekeeping and development. Then, we will discuss the current situation of China’s participation in peacekeeping operations and how data-driven methods can help China play a greater role in peacekeeping tasks. Finally, we will propose a sustainable peace and development model that combines data-driven methods with the advantages of China’s peacekeeping efforts to achieve long-term peace and development goals.

show abstract

Section: The Achievements Of Data-driven Methodologies In Several Dom...mentioning

confidence: 99%

Building a sustainable peace and development model through data-driven Chinese peacekeeping actions based on UN Global Pulse

Yan,

Zhang

2024

J. Infras. Policy. Dev.

View full text Add to dashboard Cite

show abstract

“…According to the publication authors, both methods, GMN and SGMN, achieve good prediction results in accuracy and performance. Among the most recent publications, a very interesting one seems to be 32 , whose authors present a new approach to this problem, i.e., traffic prediction for incomplete data. Also worth mentioning is publication 33 , which provides a comprehensive review and comparison of traffic prediction methods used.…”

Section: Related Workmentioning

confidence: 99%

Road traffic can be predicted by machine learning equally effectively as by complex microscopic model

Sroczyński,

Czyżewski

2023

Sci Rep

View full text Add to dashboard Cite

Since high-quality real data acquired from selected road sections are not always available, a traffic control solution can use data from software traffic simulators working offline. The results show that in contrast to microscopic traffic simulation, the algorithms employing neural networks can work in real-time, so they can be used, among others, to determine the speed displayed on variable message road signs. This paper describes an experiment to develop and test machine learning models, i.e., long short-term memory, gated recurrent unit recurrent networks, and stacked autoencoder networks. It compares their effectiveness with traffic prediction results generated using a widely recognized traffic simulator that analyzes traffic at the level of individual vehicles.

show abstract

“…Federated Learning (FL) is considered a priority solution to the above problems encountered in the VLD scenario and has been applied to areas such as mobile devices 15 , healthcare 16 , and the industrial IoT 17 , but has not yet been introduced to the VLD. FL is a new distributed training framework that uses local data to train when building networks, not sharing data, and integrates model weights from different devices to obtain better global models, with the main goal of breaking through the problem of data islands on the basis of protecting users' privacy 15 . In the direction of mobile devices, optimization of keyboard prediction 18 and action trajectory prediction 15 models by FL methods has been of great interest to researchers.…”

Section: Related Workmentioning

confidence: 99%

“…FL is a new distributed training framework that uses local data to train when building networks, not sharing data, and integrates model weights from different devices to obtain better global models, with the main goal of breaking through the problem of data islands on the basis of protecting users' privacy 15 . In the direction of mobile devices, optimization of keyboard prediction 18 and action trajectory prediction 15 models by FL methods has been of great interest to researchers. In compliance with the demand for using decentralized data to gain benefits in the industrial engineering field, FL has been widely promoted here.…”

Section: Related Workmentioning

confidence: 99%

FL-VLD: A Novel Voice Liveness Detection Framework Based on Asynchronous Federated Learning

Zong,

Li,

Zhu

et al. 2023

Preprint

View full text Add to dashboard Cite

Voice Liveness Detection (VLD) has become one of the hot research topics in the Internet of Things era. Conventional VLD methods are centralized solutions that are trained on abundant data collected from local clients and stored on the server, however, they have the risk of causing data islands and privacy leakage. To address this problem, we propose a novel word-level VLD framework based on asynchronous federated learning (FL) with pop noise, named FL-VLD. Structurally, FL-VLD takes the preprocessed voice for local model training and constructs a global model by only transmitting the learned weights after differential privacy with FL’s central server in an asynchronous manner. In addition, the local network of the framework incorporates the residual network and the spatial grouping enhancement module to optimize the complexity and accuracy of the global model. With the advantage of FL’s distributed structure, FL-VLD solves the data island problem in VLD scenario without threatening users’ privacy. Experimental results on the popular POCO dataset show that our proposal is clearly superior to the traditional centralized methods as well as overperforming other federated schemes in terms of fairness, stability, accuracy, and lightweightness. Further, for attacks involving far-field replay, synthesis, and conversion, FL-VLD has high generalisation capabilities. Finally, the ablation study attests to its efficacy.

show abstract

Traffic Prediction With Missing Data: A Multi-Task Learning Approach

Cited by 15 publications

References 35 publications

Building a sustainable peace and development model through data-driven Chinese peacekeeping actions based on UN Global Pulse

Building a sustainable peace and development model through data-driven Chinese peacekeeping actions based on UN Global Pulse

Road traffic can be predicted by machine learning equally effectively as by complex microscopic model

FL-VLD: A Novel Voice Liveness Detection Framework Based on Asynchronous Federated Learning

Contact Info

Product

Resources

About