[w3ctag/design-reviews] Web Machine Learning Model Loader API (Issue #759)

Wotcher TAG!

I'm requesting a TAG review of Web Machine Learning (WebML) Model Loader API.

The WebML Model Loader API is a proposed web API to take a custom, pre-trained machine learning (ML) model in a standard format, and apply it to some data in JavaScript to perform inference, like classification, regression, or ranking. The idea is to make it easy and performant to use a custom, pre-built machine learning model in web apps, across devices and browsers.

  - Explainer: [explainer.md](https://github.com/webmachinelearning/model-loader/blob/main/explainer.md)
  - User research:
  - Security and Privacy self-review²: [security-privacy-self-review.md](https://github.com/webmachinelearning/model-loader/blob/main/security-privacy-self-review.md)
  - GitHub repo: [webmachinelearning/model-loader](https://github.com/webmachinelearning/model-loader)
  - Primary contacts:
      - Jiewei Qian, @wacky6, Google
      - Honglin Yu, @yuhonglin, Google
      - Jonathan Bingham, @jbingham, Google
  - Organization/project driving the design: Google, in collaboration with [Machine Learning for the Web Community Group](https://webmachinelearning.github.io/)
  - External status/issue trackers for this feature:
    - [GitHub issues](https://github.com/webmachinelearning/model-loader/issues)
    - [Chrome Status](https://chromestatus.com/feature/6505078216196096)
    - [Chromium issue tracker](https://bugs.chromium.org/p/chromium/issues/list?q=component%3ABlink%3EWebML&can=2)

Further details:

  - [x] I have reviewed the TAG's [Web Platform Design Principles](https://www.w3.org/TR/design-principles/)
  - The group where the incubation/design work on this is being done (or is intended to be done in the future): [Machine Learning for the Web Community Group](https://webmachinelearning.github.io/)
  - The group where standardization of this work is intended to be done ("unknown" if not known): Same as above
  - Existing major pieces of multi-stakeholder review or discussion of this design: 
     - [Web Machine Learning WG archive](https://lists.w3.org/Archives/Public/public-webmachinelearning-wg/)
  - Major unresolved issues with or opposition to this design:
     - Standardized model format: https://github.com/webmachinelearning/model-loader/issues/28

     - GPU / accelerator support and assessing their fingerprint potential
  - [Some performance benchmark of this API vs TF.js WebAssembly](
https://docs.google.com/presentation/d/1C9s2auF32T28ckTCAZ3Yd-wxSz2syOlEIYtMk57sgR8/edit?usp=sharing)
  - This work is being funded by: N/A

We'd prefer the TAG provide feedback as (please delete all but the desired option):
  🐛 open issues in our GitHub repo for **each point of feedback**


-- 
Reply to this email directly or view it on GitHub:
https://github.com/w3ctag/design-reviews/issues/759

You are receiving this because you are subscribed to this thread.

Message ID: <w3ctag/design-reviews/issues/759@github.com>