Spring 2018 Projects with the
Smart-X {Cities, Buildings, Grids} Group

1. Deep Learning and Model Blending for Nowcasting:
Improve short-term (nowcasting) weather forecasting performance using machine learning for solar and wind energy applications.

2. Deep Learning for Intelligent Building Management.
Following up on the outstanding results from the Fall 2017 semester, we are further developing the total Property Optimizer.  We are looking for students with deep (reinforcement) learning based time series forecasting and control experience.  Experience with GANs.  Also, experience in building management systems and SCADA would be of value.

3. Causal Inference for Treatments of Con Ed assets.
We are studying the effectiveness of several inspection/maintenance programs on the Con Ed secondary grid using causal inference. Looking for statistics and causal inference experience.

4. Fast Reinforcement Learning via Quantum Computing.
We are studying ways to use quantum computing to speed up dynamic programming/reinforcement learning. Looking for people, ideally with a physics background, knowledgeable in quantum computing and (deep) reinforcement learning.

5. Deep Learning of Cascading Failures in Power Grids.
This is an ongoing project with Daniel Bienstock using his cascading failure model to train deep networks on cascading power grid data and then use deep reinforcement learning to avoid or minimize threats from these cascades. Looking for people knowledgeable in deep (reinforcement) learning and simulation of complex systems.

6. Human-in-the-loop alarm fatigue reduction
Research if false alarms in intensive care units (ICUs) can be reduced by building ML models of system/patient based on electronic health records, monitor data, and physiological data. Have unlabeled data from a neurological ICU. Looking for students to build interactive tools to analyze and visualize the data using D3.js or Microsoft Power BI Desktop and R or Python.

Contact:
Albert Boulanger
Interchurch Center, 475 Riverside Dr, Suite 850
[email protected]
212 870 1276