NEWLSTM: An Optimized Long Short-Term Memory Language Model for Sequence Prediction

Wang, Qing; Peng, Rong-Qun; Wang, Jiaqiang; Li, Zhi; Qu, Hanbing

doi:10.1109/access.2020.2985418

Cited by 19 publications

(9 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similar to all recurrent neural network-based methods, the MLP-LSTM [39] still suffers from long-term dependencies. [40][41][42] Therefore, by extracting low-quality temporal features, the performance of MLP-LSTM [39] was degraded. However, the HAMA achieved the highest performance by using the attention mechanism to extract high-quality temporal features.…”

Section: Performance Comparisonmentioning

confidence: 99%

Multimodal Neural Network for Recurrence Prediction of Papillary Thyroid Carcinoma

Kim

Lee

Choi

et al. 2023

Advanced Intelligent Systems

View full text Add to dashboard Cite

Thyroid carcinoma is the most common endocrine carcinoma, constituting approximately 3% of all carcinoma cases diagnosed worldwide as of 2020. [1] In particular, the papillary thyroid carcinoma (PTC) accounts for 84-88% of all thyroid carcinoma cases. [2][3][4][5][6] According to the surveillance, epidemiology, and end results (SEER) program, [1,6,7] the incidence and recurrence rates of PTC have steadily increased in the past decade. Some studies reported that recurrence occurred in approximately 8-28% of PTC patients, [3,8,9] which indicates that PTC frequently invades the perithyroid tissues and cervical lymph nodes (LNs). [3,6,8,9] For this reason, PTC patients require additional examinations and operations, which results in extensive morbidity and medical costs compared to those of the first operation. Therefore, early prediction of PTC recurrence can prevent the spread of serious carcinoma and reduce the socioeconomic burden on patients. [5] In the medical field, various clinical indicators have been suggested as predictors of PTC recurrence. According to the risk of structural disease recurrence of the 2015 American Thyroid Association (ATA) guidelines, the histological types of tumor, tumor size, tumor multiplicity, extrathyroidal extension (ETE), extranodal extension (ENE), and mutation of genes (BRAF, TERT) are crucial for estimating PTC recurrence. [10] Moreover, some studies considered additional features such as the demographics, thyroid function tests (TFTs), type of surgery, and I-131 radioactive therapy. [9][10][11][12][13][14][15] However, their goal was to individually analyze the association between PTC recurrence and each feature. Since no unified predictor of PTC recurrence exists, [16][17][18][19] it is necessary to develop an artificial intelligence (AI) model that can analyze these features in an integrated manner, similar to clinicians who make decisions based on diverse examinations.Recently, research on predicting thyroid carcinoma recurrence using machine learning (ML) and deep learning (DL) has been increasing. Zhu et al. [20] developed an online recurrence risk calculator based on the extreme gradient boosting (XGBoost) algorithm. Park et al. [21] categorized the number of metastatic LNs and lymph node ratio (LNR) by considering the prognostic significance of clinicopathologic features, and predicted PTC recurrence through ML algorithms using these features. To predict thyroid carcinoma recurrence using computed tomography

show abstract

Section: Performance Comparisonmentioning

confidence: 99%

Multimodal Neural Network for Recurrence Prediction of Papillary Thyroid Carcinoma

Kim

Lee

Choi

et al. 2023

Advanced Intelligent Systems

View full text Add to dashboard Cite

show abstract

“…A recursive neural network (RNN) was trained using the same training dataset and compared to ANN. The RNN can easily map sequences to sequences whenever the alignment between the inputs and outputs is known ahead of time [33]. Identically RNN with 12 input layers, 12 output layers, and 12 hidden layers was trained.…”

Section: B Machine Learning For Regression Problemmentioning

confidence: 99%

Passing Vehicle Road Occupancy Detection Using the Magnetic Sensor Array

et al. 2023

View full text Add to dashboard Cite

The increasing presence of vehicles on roads necessitates intelligent traffic management solutions in areas where video cameras cannot be utilized. Currently, there are limited choices for depersonalized vehicle reidentification systems. This paper introduces a system that later will be used for vehicle reidentification. The system uses anisotropic magnetoresistive sensors and is based on the hypothesis that each vehicle leaves unique magnetic signatures which can be used for comparison and matching. Vehicle location on the road perpendicular to sensor array detection methodology is presented in this work. An array of magnetic sensors is installed in asphalt across the vehicle's driving direction. The system continuously measures Earth's natural magnetic field and detects distortions when vehicles pass an array of a sensors and then logs magnetic signatures. Useful parameters from raw sensor axes are calculated -modules and derivatives. Applying signal-to-noise ratio calculation for module derivatives between ambient noise and signal gives important features for neural network input. Different types of neural network architectures and output result interpretation techniques are investigated. Further, after evaluating network output it is possible to label sensor nodes that are directly beneath the vehicle. Experiment results show that implemented algorithm is highly sufficient for valid sensors under the vehicle selection. Correct sensor selection is important for further re-identification algorithms.INDEX TERMS Magnetic field measurement, magnetic signature, vehicle re-identification, intelligent transportation systems.

show abstract

“…One of them is referred to as a keyframe, so one timestep will have several time series analysis items into the RNN at the same period. LSTM uses gate structures to selectively control information, thereby alleviating the defect of the disappearance gradient of traditional RNN (Wang et al, 2020). Among them, the input gate and output gate control input and output of the unit respectively, and forget the state of the gate control unit.…”

Section: Long Short-term Memorymentioning

confidence: 99%

ELSTM: An improved long short‐term memory network language model for sequence learning

Wang

et al. 2022

Expert Systems

Self Cite

View full text Add to dashboard Cite

The gated structure of the long short‐term memory (LSTM) alleviates the defects of gradient disappearance and explosion in the recurrent neural network (RNN). It has received widespread attention in sequence learning such as text analysis. Although LSTM has good performance in handling remote dependencies, information loss often occurs in long‐distance transmission. We propose a new model called ELSTM based on the computational complexity and gradient dispersion in the traditional LSTM model. This model simplifies the input gate of LSTM, reduces some time complexity by reducing some components, and improves the output gate. By introducing the exponential linear unit activation layer, the problem of gradient dispersion is alleviated. Comparing the new model with multiple existing models, when predicting language sequences, the time used by the model has been greatly reduced, and the language confusion has been reduced, showing good performance.

show abstract

NEWLSTM: An Optimized Long Short-Term Memory Language Model for Sequence Prediction

Cited by 19 publications

References 17 publications

Multimodal Neural Network for Recurrence Prediction of Papillary Thyroid Carcinoma

Multimodal Neural Network for Recurrence Prediction of Papillary Thyroid Carcinoma

Passing Vehicle Road Occupancy Detection Using the Magnetic Sensor Array

ELSTM: An improved long short‐term memory network language model for sequence learning

Contact Info

Product

Resources

About