IEEE

Autoencoder–LSTM and Graph-Theory Hybrid Framework for Secure Clustering and Lifetime Maximization in WSNs

Abstract: Smart sensors that are smaller, cheaper, and more energy efficient have been made possible by the fast evolution of MEMS technology, which has greatly increased the importance of WSN.
GitHub

striim-labs/lstm-autoencoder-spark-kafka

Real-time anomaly detection on NYC taxi demand data using an LSTM Encoder-Decoder model, Kafka streaming, Spark Structured Streaming, and a Dash visualization dashboard. Based on Malhotra et al. (2016 ...
GitHub

Space Domain Awareness AI System

A production-grade space tracking and threat assessment system that processes 1,000 space objects in real time. The system ingests TLE orbital data, runs it through a 7-model ML ensemble (4 neural + 3 ...
Frontiers

Subtyping insomnia disorder with a population graph attention autoencoder: revealing two distinct biotypes

Insomnia disorder (ID) is neurobiologically heterogeneous and often eludes characterization by traditional group-level neuroimaging. Subtyping based on neuroimaging and clinical data offers a ...
Scientific Research Publishing

A Hybrid CNN-LSTM Variational Autoencoder for Treatment Response Prediction in Synthetic Psychiatric Data ()

Treatment response prediction remains one of the most pressing challenges in precision psychiatry, where patient heterogeneity and complex biomarker interactions limit the reliability of conventional ...
IEEE

Lightweight Real-Time Anomaly Detection and Prediction for Surveillance Videos Using a Self-Supervised Hybrid Autoencoder and Dynamic Graph CNN-LSTM Framework

Abstract: Anomaly detection in surveillance videos is a critical task for ensuring public safety, requiring both high accuracy and real-time efficiency. This study presents a novel lightweight anomaly ...
Scientific Research Publishing

Graph Neural Networks for Anomaly Detection in Cloud Infrastructure ()

Cloud infrastructure anomalies cause significant downtime and financial losses (estimated at $2.5 M/hour for major services). Traditional anomaly detection methods fail to capture complex dependencies ...