Features deferred to WebNN v2

[anssiko]

(Related to the WebML WG Charter in development at #19)

The WG has deferred some WebNN API features that it believes are too early for Candidate Recommendation to "v2". We should review these v2 issues and see if they warrant explicit mention in the next charter.

The latest WebNN v2 issues are:
https://github.com/webmachinelearning/webnn/labels/v2

[anssiko]

The expectation is the WG will continue to work on WebNN v2 features under its next Charter for 2023-2025 and will continue to prioritize features that can receive implementation experience from multiple backends/libraries and exercise multiple device types. For example, XNNPACK and DirectML are two different backends for the Chromium implementation of the WebNN API. No changes to the charter scope suggested for this issue.