User Tools

Site Tools


research_hack_day_notes

Music Research Hack Day Notes

  • Less organization up front
  • Institutions that put money towards travel grants as prizes - so that you can finish something after the fact. Usual bounds of travel money. There are sources for money like that.
  • Out-of-domain people - get ML people to come in and add additional interest.
  • Too constraining to have an entire day on one task.
  • Taking people out of their comfort zone
  • Is there enough time to make something? What is the scope?
  • Dedicated period at the end for writing. Code black out time.
  • Require people to write something up because it makes it more legit.
  • Good indication of where research
  • Hit the ground running
  • It is feasible to develop a cool idea into a one page paper in a weekend.
  • How much experience is necessary? People of various experience levels could benefit.
  • People with more experience could help out younger people get experience, while getting people to work on their project.
  • Not enough people at MHD there to help other people.
  • Maintain a tinkerer mentality - research playtime
  • Have mentors around - eg professors.
  • Allow people to try something outside of their area
  • Starting point rather than a finishing point - have a conversation with someone and have a real paper come out later
  • Have an abstract not necessarily results - or have simple prizes
  • If you incentivize stuff then sometimes you can lose creativity
  • Don't want it to be just a traditional hack day but for research
  • May need to offer travel grants for certain people (who are far away) - it's a lot of effort to get to a place, and potentially expensive.
  • Look for institutions who are happy to fund something which is really different
  • NEMISIG connection, ISMIR connection would help a lot.
  • Goes along with the current trend in new research directions
  • Topic-wise good to focus on music, but good to invite people from all kinds of labs
  • Consistency with the NEMISIG branding is good - email list, websites, 5 year history
  • Figuring out intention
  • Figuring out feasibility of hosting it (getting money estimate)
  • Doing it at a place where there is internet infrastructure - may need to buy additional stuff
  • Can't really count on it in the summer
  • Get collaborative feedback early on
research_hack_day_notes.txt ยท Last modified: 2015/12/17 21:59 (external edit)