TAG Meeting 2023-09-21

TAG Meeting 2023-09-21

Agenda

  • TAG meeting frequency: should keep monthly or alter to bimonthly or even quarterly?

  • Error handling improvements: high level overview

  • Open conversation on AI, data interoperability, and the roles the Ed-Fi Alliance and the Ed-Fi Community

Materials

Participants

First Name

Last Name

Organization

Josh

Allen

Denver Public Schools

Dirk

Bradley

Michigan Datahub

Kathleen

Browning

Ed-Fi Alliance

David

Clements

Ed-Fi Alliance

Wyatt

Cothran

South Carolina Department of Education

Katie

Favara

Texas Region 4

Jean-Francois

Guertin

EdWire

Jason

Hoekstra

Ed-Fi Alliance

Manuel

Jaramillo

Texas Education Agency

Sherod

Keen

Keen Logic

Vinaya

Mayya

Ed-Fi Alliance

Mike

Minuto

Ed-Fi Alliance

Nadu

Nair

Walla Walla Pubic Schools

Oscar

Ortega

Edupoint

Ron

Peasha

Infinite Campus

Andrew

Rice

Education Analytics

Lucy

Saura

Lake Washington School District

Ann

Su

Ed-Fi Alliance (Contractor)

Rick

Thompson

South Carolina Department of Education

Maureen

Wentworth

Ed-Fi Alliance

Mustafa

Yilmaz

Ed-Fi Alliance

Support: Ann Su

Notes

Meeting Frequency

  • SIS and MSP special interest groups don't have agency representation.

  • Not seeing input to agenda directly from the TAG members - continue to solicit.

  • No clear consensus on timing. Many open to continuing with monthly, others fine with changing to bimonthly.

Error Handling

  • Question: will additional error messaging take up more log space? Answer: no change to what is output to log file.

  • Can the error messages themselves be more user friendly?

    • SF editorial response: Difficult, because Ed-Fi API language perforce won't match source system language. The model terms and names are already "universal", and specific end-user instructions will depend on the software.

AI Conversation

  • Responsible AI can drive demand for interoperability.

  • AI requires more data - both breadth (for interesting conclusions) and depth (for informative / less-biased conclusions). Data Standard makes that exchange more feasible.

  • Greatest value of AI is in generating insights, and helping people ask more question of the data.

  • Unclear if there is a role for AI in data validation.

  • IEP: yes, should provide support for unstructured text.

    • A resource for storing blobs.

    • A resource for storing summaries or decomposition of those blobs.

  • Provenance: yes, need to store that.

Next Meeting

Nov 16, 2023 members: please send any agenda suggestions.

Action Items

  • Stephen's decision for now: switch meeting schedule to bi-monthly. 

  • Ed-Fi Alliance team will continue looking into how to position the Ed-Fi Data Standard and tech stack to support future AI initiatives.