Presented By O'Reilly and Cloudera
Make Data Work
Dec 4–5, 2017: Training
Dec 5–7, 2017: Tutorials & Conference

Energy monitoring with a self-taught deep network

Yiqun Hu (Singapore Power)
5:05pm5:45pm Wednesday, December 6, 2017
Machine Learning
Location: Summit 2
Average rating: ****.
(4.83, 6 ratings)

Who is this presentation for?

  • Data scientists, machine learning engineers, and those working in energy or utilities

Prerequisite knowledge

  • A basic understanding of machine learning and deep learning

What you'll learn

  • Learn how to combine unlabeled data and labeled data for time series analysis


Energy usage is a significant part of daily life, so the ability to monitor this use offers a number of benefits, from cost savings to improved safety. A key challenge is the lack of labeled data. Yiqun Hu shares a new solution: a RNN-based network trained to learn good features from unlabeled data. The self-learned feature extractor can then be used for different downstream applications, such as appliance detection (supervised) and anomaly detection (unsupervised).

Photo of Yiqun Hu

Yiqun Hu

Singapore Power

Yiqun Hu is the head of data for SP Digital, where he is responsible for leading the data team on the development of machine learning capabilities for energy and utility applications. Previously, he helped several organizations build data-driven products such as image recognition systems and recommendation engines. Yiqun is the author of 30+ scientific publications in the machine learning area, with over 1,400 citations. He holds a PhD from Nanyang Technological University and a bachelor’s degree in computer science from Xiamen University.