| Mon 2024-12-09 | |
| 09:30-12:00 | Artificial Intelligence PTF |
| | Callin Info: | Callin access through meet.goto.com will be available to registered attendees during the event. |
Important notes- Times are listed in the PST time zone (UTC-8) since the in-person meeting is in San Francisco.
- You must register with OMG, even if you're asking to attend remotely, to have access to the collaboration platform. The GoTo link will be visible once logged in to the Guidebook app or web page.
- Contact [email protected] with any other logistics or registration questions.
- E-mail [email protected] to suggest/request additional topics or changes to the agenda.
AGENDA
Note: this is tentative as of early November and items may still be moved around.
9:30-9:50 Welcome, introductions, summary of September meeting
9:50-10:30 Neural Network Model RFP- Presentation of the draft RFP (Konstantin Oppl)
- Discussion
- The RFP will also be presented to the Architecture Board, and revisions are expected between this meeting and Feb. 17 (four-week deadline for the Q1 meeting), followed by a vote to recommend issuance in March.
10:30-10:45 Coffee break
10:45-11:15 Cloud Working Group discussion paper on "AI and the Cloud"
- Vote to recommend publication (assuming the paper is in final draft form)
- The MARS PTF will also be asked to vote as it has traditionally be the subgroup to which Cloud WG papers were submitted
11:15-11:30 RFI on needed AI standards
- Review of the status of the NIST 2019 RFI responses (needed before creating a new RFP)
11:30-12:00 "How Decentralized Identity and AI can help each other" (Markus Sabadello, Danube Tech)- Abstract: In this presentation,we explore mutual synergies between decentralized digital identity technologiesand AI. This includes how we can add provenance and trust to input data for AI models,how personal digital identity can enable personal AIs, and how digital identitytechnologies can help distinguish who is real online.
|
| Tue 2024-12-10 | |
| 09:00-12:00 | AI PTF (continued) |
| 9:00-9:15 Recap of Day 1, review of Day 2 agenda
9:15-10:00 Keynote: "AI Evaluation and Governance" -- Neel Guha, Stanford University (https://neelguha.github.io/)
- Abstract: Evaluation" is the process by which AIdevelopers produce information about the benefits and risks of AI systems. Itis a specialized technical practice–much in the same way that running crashtests or clinical trials is. In this talk, I'll describe the different forms ofevaluation that occur, the types of information they produce, and theconstraints that developers operate under when performing evaluation. I'll thentalk about the connections between evaluation and AI governance–how dodifferent aspects of evaluation shed light on questions of AI liability andregulatory design?
10:00-10:30 Open slot, or begin roadmap discussion
10:30-10:45 Break
10:45-12:00 AI PTF Roadmap
- Confirm neural network model RFP schedule (report on feedback from Architecture Board and required actions)
- List of suggested activities (Isabell Kunst)
- Discussion
- Agenda for March 2025 meeting
- Neural Network Model RFP issuance
- RFI on AI Standards
- AI Monetization paper (separated from paper on AI and the Cloud)
- (Postponed from December) "Will AI Innovation Migrate to Chip Design? Impact on Standardization" -- Dr. Karl Gosejacob
|
| 13:00-14:30 | Joint AI/C4I/ADMTF session on AI BOM |
|
On Tuesday afternoon (to be confirmed), we are planning a joint meeting with the C41 (Defense & Military) DTF and the Architecture-Driven Modernization PTF about the requirements for an AI Bill of Materials and how those are addressed by the SPDX 3.0 specification for Software Bills of Materials.
The specific agenda is in preparation. Potential topics and speakers:- The CISA SBOM program (Dr. Allan Friedman?)
- AIBOM requirements (Daniel Bardenstein or Marc Frankel, Manifest Cyber?)
- How SPDX 3.0 addresses AIBOM requirements (Bob Martin, MITRE?)
- Discussion
Meeting in room hosted by: C4IDTF
|