site templates free download

Projects

Here I present some projects in which I've been involved.

Mobirise

MSc Thesis

Electrical and Computer Engineering

Mobirise

Energy

Applied research on deep learning for time series and anomaly detection in energy data.

Mobirise

Healthcare

Applied research on deep learning for representation learning of healthcare time series data.

Mobirise

MSc Thesis

Unsupervised Anomaly Detection in Time Series Data using Deep Learning
PDF & Presentation.

Mobirise

Energy

Anomaly Detection in Solar Energy Time Series Data

This project consisted on applying deep learning to time series anomaly detection.
The proposed approach is based on a Deep Generative Model - the Variational Autoencoder - and integrates recurrent neural networks and an attention mechanism.

  1. UNSUPERVISED - no need for big labelled datasets. Just takes in the raw data and it learns everything from scratch.
  2. GENERIC - Suitable for every type of time series data (seasonal, non-seasonal, predictable, unpredictable.). Can also be applied to other sequential data structures, such as text and videos.
  3. SCALABLE - Inference and the computation of the anomaly scores are efficient, both taking in total a few milliseconds.

Conference PaperUnsupervised Anomaly Detection in Energy Time Series Data using Variational Recurrent Autoencoders with Attention

Oral Presentation in the 17th IEEE International Conference on Machine Learning and Applications (ICMLA'18).

PDF (accepted version)

Mobirise

Healthcare

Electrocardiogram Monitoring

Conference Paper: Learning Representations from Healthcare Time Series Data for Unsupervised Anomaly Detection
To appear as an Oral Presentation at the 6th IEEE International Conference on Big Data and Smart Computing (BigComp'19).
PDF (accepted version)