Zoom Logo

T3 Network PP9: Data Collaboratives for Individual-Level Data Kick-Off Webinar - Shared screen with speaker view
Matt Gee
16:22
Hey all, if you could intro yourselves in the chat, that would be great!
Dale Allen
17:09
Hi folks. Dale Allen with DXtera here. Can only stay for teh early part. Go team.
Kausar Samli
17:20
“Hi” - Kausar Samli (Learning Machine Technologies)
Patrick Browder
17:25
Hi All - This is Patrick Browder from SOLID
Rick Skeel
18:33
HI Folks. Rick Skeel with Ellucian and the PESC Board
Michael Sessa PESC
18:34
hi all.
Suria Venkat
18:50
Hey all, This is Suria Venkat from National Student Clearinghouse.
cst.jeor
19:04
Hi all! This is Chris St. Jeor from Zencos.
Tom Green
32:23
SPEEDE would probably be another example here. Rick would you agree?
Suria Venkat
34:43
Does Learner Record Index candidate for colloboration as well ?
Michael Sessa PESC
37:05
possible suggestion: plug and play model for business interoperability
Matt Gee
38:33
@suria - really good point. A Learner Record Index is definitely a data collaborative use case. It would be great if you and others could help surface LRI-specific challenges as part of the work of the group.
Suria Venkat
38:34
Also Does benefit lists assume that we will deal with de-identified user/learner/student information?
Matt Gee
40:46
@Suria - I think that is a critical issue for guidance: for what data collaborative use cases are de-identified data essential vs not. I think we should generally assume that most data collaboratives will need to do their pooled analytics on de-identified data. This is especially true if you want to, for example, expand access to the data for more researchers.
Tom Green
41:19
@suria - I was wondering about this, too. Could one goal be that we have great access to data but that it maintains the privacy and security of learner-level data? Are those mutually exclusive or is it possible?
Matt Gee
41:34
@Michael - would love to hear more about your ideas on plug-and-play model for business interoperability. What are you thinking?
Suria Venkat
42:08
@Matt, @Suria - Those calls for detailed use case deep dive, may be a sub-topic for later agenda.
Matt Gee
42:29
Here’s the link for adding shared technical challenges: https://docs.google.com/spreadsheets/d/1QKhHGLYrro5FS93GkfwDxKEmDZQ_wGnvaILRt7Ci5WI/edit?usp=sharing
Michael Sessa PESC
42:52
faster understanding of how the ecosystem works, a 'single point of contact', clear and available standards, common understanding on the value of data, and what data = faster, better business services
Michael Sessa PESC
43:10
reference big data
Matt Gee
43:22
Ah, really good points. Thanks @Michael
Tom Green
47:05
Most times, the only exceptions to student release are in emergencies and crimes! I don’t think we are heading toward those.
Jess Stahl
47:25
De-identification does not preserve privacy because it can usually be fairly easily re-identified. What about differential privacy and federated learning?
Rick Skeel
48:51
Tom Greene - Electronic Transcripts, which the SPEEDE Committee has long supported, is not an aggregated data colabortive. It is more a direct exchange from one entity to another. The individual schools maintain the data, but no one else.
Tom Green
49:21
@Rick - thanks, that is helpful
Jess Stahl
51:08
I fyou use federated learning, IRBs should be much more comfortable
Michael Sessa PESC
52:20
enforcement and/or authority...always challenging
Matt Gee
55:16
Here’s the T3 Slack link: https://bit.ly/2kAkG65
Michael Sessa PESC
55:26
big question: how is the output of this group to be used, where, when, with whom
Michael Sessa PESC
01:00:27
yes, great, thank you!
Kausar Samli
01:02:48
Thank you!
cst.jeor
01:02:53
Thanks you!