FilterNet: A Many-to-Many Deep Learning Architecture for Time Series Classification

Chambers, Robert; Yoder, Nathanael C.

doi:10.20944/preprints202002.0318.v2

Cited by 2 publications

(6 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, there is no significant improvement once the batch size passes the 64-mark. The drop in performance as window length increases may not have come as a surprise as some authors have suggested that models containing LSTM layers suffer from poor parallelism across longer window lengths [61]. Although we did not investigate further, Chambers and Yoder [61] suggest that increasing the LSTM layers can enhance the model's capacity to learn longer window lengths.…”

Section: Tuning For Window Size Segmentationmentioning

confidence: 81%

“…The drop in performance as window length increases may not have come as a surprise as some authors have suggested that models containing LSTM layers suffer from poor parallelism across longer window lengths [61]. Although we did not investigate further, Chambers and Yoder [61] suggest that increasing the LSTM layers can enhance the model's capacity to learn longer window lengths. Instead, we investigated the interaction between the two parameters (Figure 13) and discovered that window length of 32 and any batch size between 64, 128 and 512 are good for the proposed model(Figures 13 & 14).…”

Section: Tuning For Window Size Segmentationmentioning

confidence: 81%

“…The shift parameter helps create the overlap between windows as seen in Figure 7 below. In the case of a 50% shift, which is commonly used in many HAR literatures, an overlap is created such that the first half of a window contains the observations from the last half of the previous window [32,47,61]. In this work, we used a window length of 32 (circa 0.75 seconds of sensor signals) and a 50% shift.…”

Section: Data Pre-processing and Segmentationmentioning

confidence: 99%

“…There are several studies that provide sufficient guidance on how to approach grid search, especially when a hybrid CNN-LSTM model is involved. [22,61]. Both works analysed the effect and shared the many benefits of adding the CNN layers before the LSTM to solve speech and activity recognition problems.…”

Section: Hyper-parameter Tuningmentioning

confidence: 99%

“…overfitting. The literature identifies two methods suitable for preventing this: early stopping and model checkpointing [61]. We invoked the Keras EarlyStopping callback and specified the validation loss as the performance metric to monitor in order to end training.…”

Section: Hyper-parameter Tuningmentioning

confidence: 99%

See 4 more Smart Citations

Context-Aware Complex Human Activity Recognition using Hybrid Deep Learning Models

Otebolaku¹,

Omolaja²

2022

Preprint

View full text Add to dashboard Cite

Smart devices such as smartphones, smartwatches, etc. are promising platforms that are being used for automatic recognition of human activities. However, it is difficult to accurately monitor complex human activities due to inter-class pattern similarity, which occurs when different human activities exhibit similar signal patterns or characteristics. Current smartphone-based recognition systems depend on the traditional sensors such as accelerometer and gyroscope, which are inbuilt in these devices. Therefore, apart from using information from the traditional sensors, these systems lack contextual information to support automatic activity recognition. In this article, we explore environment contexts such as illumination(light conditions) and noise level to support sensory data obtained from the traditional sensors using a hybrid of Convolutional Neural Networks and Long Short Time Memory(CNN_LSTM) learning models. The models performed sensor fusion by augmenting the low-level sensor signals with rich contextual data to improve recognition and generalisation ability of the proposed solution. Two sets of experiments were performed to validate the proposed solution. The first set of experiments used inertial sensing data whilst the second set of extensive experiments combined inertial signals with contextual information from environment sensing data. Obtained results demonstrate that contextual information such as environment noise level and illumination using hybrid deep learning models achieved better recognition accuracy than the traditional activity recognition models without contextual information.

show abstract

Section: Tuning For Window Size Segmentationmentioning

confidence: 81%

Section: Tuning For Window Size Segmentationmentioning

confidence: 81%

Section: Data Pre-processing and Segmentationmentioning

confidence: 99%

Section: Hyper-parameter Tuningmentioning

confidence: 99%

Section: Hyper-parameter Tuningmentioning

confidence: 99%

See 3 more Smart Citations

Context-Aware Complex Human Activity Recognition using Hybrid Deep Learning Models

Otebolaku¹,

Omolaja²

2022

Preprint

View full text Add to dashboard Cite

show abstract

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

FilterNet: A Many-to-Many Deep Learning Architecture for Time Series Classification

Cited by 2 publications

References 31 publications

Context-Aware Complex Human Activity Recognition using Hybrid Deep Learning Models

Context-Aware Complex Human Activity Recognition using Hybrid Deep Learning Models

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

Contact Info

Product

Resources

About