WebML WG Teleconference – 18 November 2021 - 15:00-16:00 UTC+0

Latest version: https://github.com/webmachinelearning/meetings/blob/main/telcons/2021-11-18-agenda.md


WebML WG Teleconference – 18 November 2021 - 15:00-16:00 UTC+0

⚠️ Note the adjusted (+1 hour) meeting start times in China and Japan due to Daylight Saving Time change<https://www.timeanddate.com/time/dst/events.html>.

See the timezone table ...
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2021-11-18-agenda.md#logistics>Logistics

  *   Chair: Anssi
  *   Scribe: ?
  *   IRC: irc://irc.w3.org:6667/#webmachinelearning

  *   IRC web client: https://irc.w3.org/?channels=#webmachinelearning

  *   Call-in details: https://lists.w3.org/Archives/Member/internal-webmachinelearning/2020Apr/0000.html

  *   Minutes: https://www.w3.org/2021/11/18-webmachinelearning-minutes.html


________________________________
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2021-11-18-agenda.md#agenda>Agenda
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2021-11-18-agenda.md#%E2%84%B9%EF%B8%8F-tpac-meeting-follow-up>ℹ️ TPAC meeting follow-up

We had a broad range of exciting discussion at TPAC with guest appearances. Thank you all for your contributions!

Let's reflect on our TPAC meeting learnings and discuss what next steps the group should take informed by these new insights and perspectives.

TPAC agenda: https://github.com/webmachinelearning/meetings/issues/18


TPAC minutes: Day 1<https://www.w3.org/2021/10/26-webmachinelearning-minutes.html>, Day 2<https://www.w3.org/2021/10/27-webmachinelearning-minutes.html>, Day 3<https://www.w3.org/2021/10/28-webmachinelearning-minutes.html>

Below are some questions and proposals that emerged from our discussions. Further suggestions are welcome.

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2021-11-18-agenda.md#%E2%84%B9%EF%B8%8F-rationalecriteria-for-adding-new-ops-to-the-webnn-api>ℹ️ Rationale/criteria for adding new ops to the WebNN API

  *   What are the key takeaways from other projects in managing evolution to inform our work? (presentation slides<https://lists.w3.org/Archives/Public/www-archive/2021Nov/att-0000/W3C_Adding_new_Operators.pdf>)
  *   Gauge interest for a light-weight process (issue template?) to guide submitting new operator requests to WebNN.

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2021-11-18-agenda.md#%E2%84%B9%EF%B8%8F-versioning-and-web-compatibility>ℹ️ Versioning and web compatibility

  *   Proposal to request incremental TAG reviews for substantial changes.

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2021-11-18-agenda.md#%E2%84%B9%EF%B8%8F-privacy-and-security-discussion>ℹ️ Privacy and security discussion

  *   Proposal to evaluate APIs against fingerprinting best practices<https://www.w3.org/TR/fingerprinting-guidance/#bp-summary> and severity<https://www.w3.org/TR/fingerprinting-guidance/#identifying-fingerprinting-surface-and-evaluating-severity>
  *   WebGPU discussion takeaways to inform the AI accelerator device selection? https://github.com/webmachinelearning/webnn/issues/169


<https://github.com/webmachinelearning/meetings/blob/main/telcons/2021-11-18-agenda.md#%E2%84%B9%EF%B8%8F-ml-js-framework-performance-focus-areas-for-webnn>ℹ️ ML JS framework performance, focus areas for WebNN

  *   Discuss API design changes, if any, informed by the performance or framework integration experience? (presentation slides<https://lists.w3.org/Archives/Public/www-archive/2021Oct/att-0014/WebNN_ML_JS_Framework_Performance.pdf>)

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2021-11-18-agenda.md#%E2%84%B9%EF%B8%8F-integrating-an-open-source-cross-platform-implementation-of-the-web-neural-network-api-into-a-web-engine>ℹ️ Integrating an open-source cross-platform implementation of the Web Neural Network API into a web engine

  *   Discuss Chromium integration next steps (presentation slides<https://lists.w3.org/Archives/Public/www-archive/2021Oct/att-0015/Integrate_WebNN-native_into_Chromium_TPAC.pdf>)
  *   Proposal to start documenting implementation status https://github.com/webmachinelearning/webnn/issues/223

  *   Call for Review of the Chromium Design Doc<https://docs.google.com/document/d/1KDVuz38fx3SpLVdE8FzCCqASjFfOBXcJWj124jP7ZZ4/>

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2021-11-18-agenda.md#%E2%84%B9%EF%B8%8F-conformance-testing-of-webnn-api>ℹ️ Conformance testing of WebNN API

  *   Discuss practical aspects of a reference implementation for the WebNN API to establish an ideal baseline.
  *   Discuss any WebNN API web-platform-tests changes informed by DirectML conformance work?

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2021-11-18-agenda.md#%E2%84%B9%EF%B8%8F-ethical-issues-in-using-machine-learning-on-the-web>ℹ️ Ethical issues in using Machine Learning on the Web

  *   Call to connect with experts in ethics to contribute to this work.