StratML for AIKR follow up notes

At the March10 online meeting we agreed to continue the work of assessing
the value of StratML for AIKR.


   - Two of the invited experts agreed to join the CG.


To facilitate our work Paola setup the AIKRREPO cloud storage

   - Audio of first two meetings are stored in  AIKRREPO

We agreed that we should Walk the Talk  and use StratML to manage our work :

   - AIKRCG will have  a Strategy to  'assess the value of StratML for AIKR'
   - AIKRCG will use StratMl to document how 'Evaluate Trustworthy AIKR
   objects implemented by machine learning powered services'


Latest draft of the document 'Evaluate Trustworthy AIKR objects implemented
by machine learning powered services' is stored in AIKRREPO

Next meeting is scheduled for Tuesday March 31st

Carl Mattocks
cochair AIKRCG

Be Safe to Keep Healthy



On Tue, Mar 10, 2020 at 9:09 PM Paola Di Maio <paola.dimaio@gmail.com>
wrote:

> Thanks to those who participated in the call yesterday, nice to hear
> people voices
> and please share follow up emails on the list perhaps?
> Look forward to the summary of actions, Carl please remind us what we are
> supposed to do
> Two quick points from me
> 1. Chris and Owen said that when they tried it the search function
> generated with google code over stratml documents didnt work. the first
> thing that comes to mind is a granularity problem.  If the GOALS in Stratml
> are chunked  as paragraphs, even long wordy paragraphs, and the search
> function/parser is set to identify individual phonemes/words, then running
> the query would not yield results, or at least not good ones
> Could it be a question of modelling the documents in a more granular way,
> or simply attaching more specific metadata/keywords to each set - say,
> GOALS so that the engines can easily identify matching content? Assume we
> can add keywords to each section of the strategy plan then these keywords
> would also have to be added to the search engine (tell the engine what to
> look for using certain parameters as opposed to guessing). This could
> increase the precision of the recall
> Let me know if this make sense?  Glad to see that jorge is keen to work on
> the query, look forward
> to be seeing what he his planning
>
> 2.  is there a workflow/dependency between
> GOAL(s), ACTION(s), OUTCOME RESULT   RECOMMENDATION.more action in StratML?
>
> I have seen in a related post call email that it is possible to set up
> reminders which I think would do the job, but it would be also good imho to
> model GOAL ACTION OUTCOME RESULT new action as a set
> Or something similar
>
> Will reply to off list emails soon, thanks a lot for helping to advance
> this
> I encourage off ist emails to be shared on list mostly for archiving
> purposes
>
> Thanks
>
> PDM
>