6_Machine Learning

  1. GAN

  2. Reinforcement Learning


  3. Usefull Resources

2. Data Types:

  • Weather data

  • Market data

  • SCADA data, PMU Data, smart meter data

  • networks data

  • outage data

  • customer data

  • GIS data

3. Research Needs

  • Data storage, access, and management- network compute

  • Data Science platform (secure, privacy, metadata & data management)b

  • visualization tool for reporting capability

    • Data policy and access

  • Data Science R&D lifecycle to make the use repeatable

  • Acquire - Store - Clean - report -Analyze

  • artificial intelligence: recognize the values in the data

  • Synthetic Data Sets

  • communication network data storage (secure and fast to retrieve, data management, processing power,

    • Goal: better and informative decisions

    • challenges: volume, variety, velocity, not structured data, have different qualities and integrity levels, can be bad or mission or outliers,

    • Techniques: determine the relevancy,

    • Applications: customized DR, asset management, EV/PV/Wind integration,

    • Should we use only real data? In other domain, data can be simulated data. They are also informative.

4. Tools and environment

    • HPC Environment (Linux Cluster), Numerous MS SOL instances, My SQL, MS Access, XLS, CSV, AWS data store

    • SQL, SAS, SPSS, Python, Matlab

5. Related Discussions