Meeting 25 June 2014, Embecosm

MAGEEC Meeting at Lymington, 25th June 2014


Present: George, Greg, Stoil, Craig, Oliver, Simon H, Simon C, James P, Jeremy


To Discuss


  • BEEBS V2 release
  • Case studies
  • WP5 July: Review of ML approach – i.e. test the previous choice with BEEBS V2
  • WP6: Training and testing (Embecosm lead; ?Does Bristol need to help?; overdue)
  • WP8: Infrastructure evaluation

[[Image:]]


Progress

  • [SC]We have a working automatic training script that can take BEEBS and pass to train the ML.
  • [SC] Working on completing milestone 6/1. Can be completed without need for Bristol interaction.

Actions and Plan forward

  • All to sign up for mailing lists
  • Plan for blog posts on each topic. Min 1 per person
  • Freenode IRC: #MAGEEC
  • BEEBS V2 to be considered by Andrew Burgess. Some weeks off a release.
  • [SH]: Request Moon if we can publish his MSc on mageec.org
  • Increase of training set
  • [SC]: Need to consider the dependencies between passes (i.e. what passes must run together) and ensure that the MAGEEC infrastructure maintains these. Is an extra constraint database necessary to express these, rather than just relying on GCC to fail.
  • [JR] Needs to put effort to complete milestone 7/2
  • Deadline for presentation of work 18th July:
    • WP6/2
      • Embecosm to lead. Discussion at next meeting of proof of concept training results.
    • WP5/3
      • Embecosm to lead.
    • WP8/1
      • This is the aggregate of work at Bristol over the Summer
    • WP8/2
      • [SH/KIE/OR] Short paper on initial results or initial draft of final journal paper.
    • WP8/3
      • [SH/KIE/OR] review targets for the final, long journal paper.

People allocation


Stoil:

  • Identification and development of Case Studies, including
    • Audio
    • Software-defined radio
  • BEEBS development
  • Add new embedded architectures
  • Energy measurement analysis
  • Community involvement
  • Remote USB power-upper-downer

George:

  • Measurement to feed-in Framework
  • Database integration
  • Expect interfaces
  • Case studies

Greg:

  • Review of ML approach
  • ML development and evaluation with Milepost Features
  • Extension of BEEBS
  • Look at PCA again once BEEBS is large enough

Craig:

  • ILP & PCA
  • Paper development

Pierre:

  • To spend time with Greg, Stoil and Craig some time next week