Transfer Learning With Time Series Data: A Systematic Mapping Study

Weber, Manuel; Auch, Maximilian; Doblander, Christoph; Mandl, Peter; Jacobsen, Hans‐Arno

doi:10.1109/access.2021.3134628

Cited by 35 publications

(19 citation statements)

References 149 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The proposed method helps avoid mismatched actions from the former single trainee activity classifier, which corrects the team training problems. The presented approach can achieve accuracy improvement compared to other methods [ 8 , 9 , 10 , 11 , 12 ]. The major contributions and key features of this paper are: Proposing a human activity recognition approach based on DNN for classification with a GAN model; Presenting a variational autoencoder (VAE) model method to denoise and extract the signal from inertial sensors in the environment; Using synthetic skeleton information to recognize the activities of individuals without using a vision approach; The proposed system can be scaled up to a ten-sensor deployment of the whole body or scaled down to two-sensor deployment of the human body; Recognizing the activities of a group or a team via pre-training model-based transfer learning.…”

Section: Introductionmentioning

confidence: 89%

A Novel Deep Neural Network Method for HAR-Based Team Training Using Body-Worn Inertial Sensors

Fan

Tseng

Wen

2022

Sensors

View full text Add to dashboard Cite

Human activity recognition (HAR) became a challenging issue in recent years. In this paper, we propose a novel approach to tackle indistinguishable activity recognition based on human wearable sensors. Generally speaking, vision-based solutions struggle with low illumination environments and partial occlusion problems. In contrast, wearable inertial sensors can tackle this problem and avoid revealing personal privacy. We address the issue by building a multistage deep neural network framework that interprets accelerometer, gyroscope, and magnetometer data that provide useful information of human activities. Initially, the stage of variational autoencoders (VAE) can extract the crucial information from raw data of inertial measurement units (IMUs). Furthermore, the stage of generative adversarial networks (GANs) can generate more realistic human activities. Finally, the transfer learning method is applied to enhance the performance of the target domain, which builds a robust and effective model to recognize human activities.

show abstract

Section: Introductionmentioning

confidence: 89%

A Novel Deep Neural Network Method for HAR-Based Team Training Using Body-Worn Inertial Sensors

Fan

Tseng

Wen

2022

Sensors

View full text Add to dashboard Cite

show abstract

“…The most extensive one focuses on the general concept of TL over all domains, including experimental results on text categorization and object detection [ 10 ]. More recently, a systematic study was published that had focused on TL in the context of time series data, discovering that the publication frequency has sharply increased since the year 2019 [ 11 ]. The study presented in [ 12 ] provides a detailed summary of TL concepts, use cases, and trends in the context of industrial automation.…”

Section: Literature Researchmentioning

confidence: 99%

Industrial Transfer Learning for Multivariate Time Series Segmentation: A Case Study on Hydraulic Pump Testing Cycles

Gaugel

Reichert

2023

Sensors

View full text Add to dashboard Cite

Industrial data scarcity is one of the largest factors holding back the widespread use of machine learning in manufacturing. To overcome this problem, the concept of transfer learning was developed and has received much attention in recent industrial research. This paper focuses on the problem of time series segmentation and presents the first in-depth research on transfer learning for deep learning-based time series segmentation on the industrial use case of end-of-line pump testing. In particular, we investigate whether the performance of deep learning models can be increased by pretraining the network with data from other domains. Three different scenarios are analyzed: source and target data being closely related, source and target data being distantly related, and source and target data being non-related. The results demonstrate that transfer learning can enhance the performance of time series segmentation models with respect to accuracy and training speed. The benefit can be most clearly seen in scenarios where source and training data are closely related and the number of target training data samples is lowest. However, in the scenario of non-related datasets, cases of negative transfer learning were observed as well. Thus, the research emphasizes the potential, but also the challenges, of industrial transfer learning.

show abstract

“…Those applications focus on systems changing over time. In a survey, Weber et al [17] reports about 223 publications concerning TL with time series data, but only one article dedicated to an industrial application of fault diagnosis is mentioned.…”

Section: State Of the Artmentioning

confidence: 99%

Adaptation of Dynamic Data‐Driven Models for Real‐Time Applications: From Simulated to Real Batch Distillation Trajectories by Transfer Learning

Rihm

Schueler

Nentwich

et al. 2023

Chemie Ingenieur Technik

View full text Add to dashboard Cite

In the absence of knowledge about challenging dynamic phenomena involved in batch distillation processes, e.g., complex flow regimes or appearing and vanishing phases, generation of accurate mechanistic models is limited. Real plant data containing this missing information is scarce, also limiting the use of data‐driven models. To exploit the information contained in measurement data and a related but inaccurate first‐principles model, transfer learning from simulated to real plant data is analyzed. For the use case of a batch distillation column, the adapted model provides more accurate predictions than a data‐driven model trained exclusively on scarce real plant data or simulated data. Its enhanced convergence and lower computational cost make it suitable for optimization in real‐time.

show abstract

Transfer Learning With Time Series Data: A Systematic Mapping Study

Cited by 35 publications

References 149 publications

A Novel Deep Neural Network Method for HAR-Based Team Training Using Body-Worn Inertial Sensors

A Novel Deep Neural Network Method for HAR-Based Team Training Using Body-Worn Inertial Sensors

Industrial Transfer Learning for Multivariate Time Series Segmentation: A Case Study on Hydraulic Pump Testing Cycles

Adaptation of Dynamic Data‐Driven Models for Real‐Time Applications: From Simulated to Real Batch Distillation Trajectories by Transfer Learning

Contact Info

Product

Resources

About