LSTM Autoencoder

LSTM autoencoders are recurrent neural networks used primarily for unsupervised anomaly detection and data compression in time-series data. Current research focuses on improving their robustness to noise, enhancing anomaly detection accuracy (often by incorporating attention mechanisms or denoising architectures), and applying them to diverse applications such as predictive maintenance, cryptocurrency fraud detection, and GPS trajectory compression. These models offer significant advantages in handling complex temporal dependencies within noisy data, leading to improved performance in various fields compared to traditional methods.

Papers

September 12, 2023

Robust-MBDL: A Robust Multi-branch Deep Learning Based Model for Remaining Useful Life Prediction and Operational Condition Identification of Rotating Machines
Khoa Tran, Hai-Canh Vu, Lam Pham, Nassim Boudaoud
Frequency Domain Useful Life Model Based Deep Learning Condition Monitoring Industrial Machine Time Domain Vibration Data Multi Branch LSTM Autoencoder

April 20, 2023

An Attention Free Conditional Autoencoder For Anomaly Detection in Cryptocurrencies
Hugo Inzirillo, Ludovic De Villelongue
Anomaly Detection Attention Based Conditional Autoencoder LSTM Autoencoder

January 18, 2023

Compression of GPS Trajectories using Autoencoders
Michael Kölle, Steffen Illium, Carsten Hahn, Lorenz Schauer, Johannes Hutter, Claudia Linnhoff-Popien
Supervised Autoencoder Linear Compression Dynamic Time Warping GPS Trajectory LSTM Autoencoder Trajectory Truncation

September 12, 2022

Detecting Driver Drowsiness as an Anomaly Using LSTM Autoencoders
Gülin Tüfekci, Alper Kayabaşi, Erdem Akagündüz, İlkay Ulusoy
LSTM Network Unlabeled Anomaly Drowsiness Detection LSTM Autoencoder Driver Drowsiness

August 30, 2022

Denoising Architecture for Unsupervised Anomaly Detection in Time-Series
Wadie Skaf, Tomáš Horváth
Anomaly Detection Long Short Term Memory LSTM Network Unsupervised Anomaly Detection LSTM Autoencoder

March 8, 2022

Visual anomaly detection in video by variational autoencoder
Faraz Waseem, Rafael Perez Martinez, Chris Wu
Supervised Autoencoder Variational Autoencoder Source Video Video Anomaly Detection Visual Anomaly Detection Convolutional LSTM LSTM Autoencoder

March 3, 2022

Comparison of LSTM autoencoder based deep learning enabled Bayesian inference using two time series reconstruction approaches
Saumik Dana
Deep Learning Consistent Comparison LSTM Network Bayesian Inference Time Series Analysis Bayesian Inference Framework LSTM Autoencoder Bidirectional Displacement