/
2023-11 - AI: LFN AI Taskforce in-person work meeting

2023-11 - AI: LFN AI Taskforce in-person work meeting

Short Description

90m, @Kenny Paul , @Casey Cain  @Lingli Deng 

Work meeting of the LFN AI Taskforce

Topic Overview

  • Overview of A General Implementation with Intent-Based Network in ONAP  work (https://youtu.be/I5g066xiXmk?si=uSjfHYBcZvfmiVUx ) - @Lingli Deng 

  • Follow up on Thoth presentation - How can the LFN community get involved in moving Thoth forward? - @Sridhar Rao 

  • Continued use case analysis - @Muddasar Ahmed 

Session Type

 

Interactive technical presentation and/or capability demo

 

Pre-Recorded technical presentation and/or capability demo

 

Live / hands-on technical training

X

Community process, governance or collaboration development

 

Real-time troubleshooting, bug-scrub or similar

 

Other (please describe)

Slides & Recording

YouTube

Following the event session recordings may be uploaded to our YouTube Channel at the full discretion of LFN staff.

Notes From Session (as appropriate)

  • Presentation by @Lingli Deng - Please upload slides →

    • Is the Open AI Platform open to the public? Yes, it is publicly accessible, although the portal is in Chinese only

    • Platform link - https://jiutian.10086.cn/open/#/firstpage?platform=OpenInnovation

    • Can the Thoth project access the open AI platform data? In theory, yes. China Mobile contributed a data set to Thoth two years ago, and now there are additional data sets

    • @Hui Deng - It could be beneficial to create an hardware abstraction layer that will enable using different types of hardware for the open AI platform. @Lingli Deng - that could be considered

    • @Muddasar Ahmed - Was OPEX savings a driver in prioritizing the use cases? @Lingli Deng - Yes, this was a consideration, practically this was not where work started, but it is interesting

  • Follow up on Thoth/Data anonymization - @rohit s

    • Seeking input on Anonymization rules

    • Call for contribution - GPU resources, interns, participation in advisory board, private data sets, testing

    • GPU and interns are the most critical right now

    • @Kenny Paul - Can Anuket's lab-as-a-service be used as hardware? @rohit s - Not enough GPU there

  • Use case prioritization matrix presented by @Muddasar Ahmed 

    • @Lingli Deng - Another column should be added to the matrix to indicate whether data sets are available. Additional implementation oriented columns.

    • @Lingli Deng - Another table for generally re-useable AI capabilities should be added to solicit interest from AI capability contributors

    • Table was added to https://lf-networking.atlassian.net/wiki/x/yHTv - everybody please comment

 

 

 

  • Action Items (as appropriate)

Use case analysis table was added, please add comments and feedback at URL below