13:11:16 RRSAgent has joined #webmachinelearning 13:11:16 logging to https://www.w3.org/2022/06/16-webmachinelearning-irc 13:11:18 Zakim has joined #webmachinelearning 13:53:35 anssik has joined #webmachinelearning 13:53:58 Zakim, prepare meeting 13:53:59 RRSAgent, make logs Public 13:54:00 please title this meeting ("meeting: ..."), anssik 13:54:10 Meeting: WebML WG Teleconference – 16 June 2022 13:54:17 Chair: Anssi 13:54:29 Agenda: https://github.com/webmachinelearning/meetings/blob/main/telcons/2022-06-16-wg-agenda.md 13:54:58 Scribe: Anssi 13:55:02 scribeNick: anssik 13:55:08 scribe+ dom 13:55:39 Present+ Anssi_Kostiainen 13:55:50 RRSAgent, draft minutes 13:55:50 I have made the request to generate https://www.w3.org/2022/06/16-webmachinelearning-minutes.html anssik 13:57:59 ghurlbot, this is webmachinelearning/webnn 13:57:59 anssik, OK 14:01:11 Present+ Ningxin_Hu 14:01:14 ningxin_hu has joined #webmachinelearning 14:01:16 Present+ Dom 14:01:29 Present+ 14:01:41 Present+ Rafael_Cintron 14:03:17 RafaelCintron has joined #webmachinelearning 14:04:23 Topic: Privacy review refresh 14:05:01 -> Previous meeting agenda https://github.com/webmachinelearning/meetings/blob/main/telcons/2022-06-02-wg-agenda.md#%E2%84%B9%EF%B8%8F-privacy-review-refresh 14:05:40 anssik: Last time we made a resolution to Request Privacy Review refresh after the WebGPU (#175 and #85) and Device Selection (#169) privacy comments have been acknowledged in Privacy Considerations 14:05:40 https://github.com/webmachinelearning/webnn/issues/175 -> Issue 175 Related WebGPU/GL Security and Privacy Considerations (anssiko) privacy-tracker, security-tracker 14:05:40 https://github.com/webmachinelearning/webnn/issues/85 -> Issue 85 Fingerprinting via machine-specific artifacts (kpu) privacy-tracker 14:05:40 https://github.com/webmachinelearning/webnn/issues/169 -> Issue 169 Device selection with MLDevicePreference and MLPowerPreference (anssiko) privacy-tracker 14:05:55 -> Resolution https://www.w3.org/2022/06/02-webmachinelearning-minutes.html#r03 14:06:09 anssik: issue #85 #175 and #169 are proposed to be addressed by PR #271 14:06:09 https://github.com/webmachinelearning/webnn/issues/271 -> Pull Request 271 Privacy Considerations refresh (anssiko) 14:07:24 anssik: ... PTAL Ningxin, Chai, Rafael, others interested in privacy aspects 14:07:31 ... questions, comments? 14:07:39 rafael: are you asking for review of #271? 14:07:51 anssik: yes - that PR aims to address by the mentioned issues 14:08:08 ... we want at least to acknowledge known privacy issues in the privacy consideration sections 14:08:17 ... in some cases we can specific normative mitigations 14:08:23 ... here it's mostly informative text 14:09:14 Topic: Ethical considerations 14:09:21 ghurlbot, this is webmachinelearning/ethical-webmachinelearning 14:09:21 anssik, OK 14:09:51 anssik: would like to discuss the plan of publishing the First Public Working Group Note of the ethical considerations document 14:09:55 anssik: most recently we merged ethics workshop feedback PR #20 14:09:55 https://github.com/webmachinelearning/ethical-webmachinelearning/issues/20 -> Pull Request 20 [closed] Incorporate feedback from the Ethical ML workshops (anssiko) 14:10:03 -> Ethical Principles for Web Machine Learning (Editor's Draft) https://webmachinelearning.github.io/ethical-webmachinelearning/ 14:10:05 ... recently we've merged #20 based on the input from the ethical ML workshops 14:10:13 -> W3C Note https://www.w3.org/2021/Process-20211102/#WGNote 14:10:22 ... the proposal is to publish the editors draft as a W3C Note 14:10:27 ... "A Group Note (NOTE) is published to provide a stable reference for a useful document that is not intended to be a formal standard." 14:10:38 ... "Some Notes are developed through successive Draft Notes before publication as a full Notes, while others are published directly as a Note. There are few formal requirements to publish a document as a Note or Draft Note, and they have no standing as a recommendation of W3C but are simply documents preserved for historical reference." 14:10:58 ... it would be an informative reference in the WebNN API 14:11:22 ... I think it is appropriate to publish a first note now, and publish updated via the automatic updates from the editors draft 14:12:14 dom: maybe we should do a CfC on the list? 14:12:18 anssik: let's do that 14:12:31 proposed RESOLUTION: Start a CfC to publish Ethical Principles for Web Machine Learning as a Draft Note and set up the document for automatic publication 14:13:05 RESOLUTION: Start a CfC to publish Ethical Principles for Web Machine Learning as a Draft Note and set up the document for automatic publication 14:13:27 Topic: Revisit WebNN API Candidate Recommendation scope 14:13:35 ghurlbot, this is webmachinelearning/webnn 14:13:35 anssik, OK 14:13:46 -> Current CR issues https://github.com/webmachinelearning/webnn/issues?q=is%3Aissue+is%3Aopen+label%3Acr 14:14:15 anssik: new "cr" issue #272, discussed later today 14:14:15 https://github.com/webmachinelearning/webnn/issues/272 -> Issue 272 Support asynchronous context creation (huningxin) cr 14:14:51 ... Review proposed approach to open use cases, proposed actions, please let us know if you have concerns: 14:14:56 ... label post-CR or close #207 14:14:56 https://github.com/webmachinelearning/webnn/issues/207 -> Pull Request 207 Update "Performance Adaptation" use case (spshin3) 14:18:06 dom: #207 should probably become an issue where the design could be further fleshed out, documenting what it'd take to support this 14:18:21 ... or if this won't be a good fit for the WebNN API, se should document that in the PR and close it 14:18:30 s/se /we / 14:19:00 anssik: my feeling is that it's clearly not for our v1 CR; maybe for future work 14:19:14 ... but Chai has been arguing this is not implementable in the first place 14:19:26 ... I'll mention this discussion in the PR 14:19:46 #253 14:19:46 https://github.com/webmachinelearning/webnn/issues/253 -> Pull Request 253 Add "Ethical Content Filtering" use case to WebNN specs (humeranoor) 14:20:17 anssik: I've suggested a few possible next steps 14:20:28 q+ 14:20:53 ack dom 14:21:33 dom: I guess on the high level since we're seeing some interest in use cases, maybe we should decouple use cases from WebNN API 14:22:25 q+ 14:23:39 dom: suggestion, take all WebNN use cases and bring them to a new doc that has use cases that WebNN API as of today satisfy and also future use cases for ML acceleration in the browser 14:23:40 q? 14:23:57 ack ningxin_hu 14:24:12 ningxin_hu: this feels like a good use case but out of scope for WebNN 14:24:14 ningxin_hu: I personally feel this is a good use case but out of scope of WebNN API 14:24:30 ... within the CG, we have the proposals repo which is another place where this could be redirected for incubation 14:24:34 ... within the CG we have the proposals repo so could use that for incubating this idea 14:24:36 https://github.com/webmachinelearning/proposals/issues 14:24:40 +1 to using the proposals repo in the CG 14:25:53 -> Current CR issues https://github.com/webmachinelearning/webnn/issues?q=is%3Aissue+is%3Aopen+label%3Acr 14:27:16 anssik: Integration with real-time video processing 14:28:24 ningxin_hu: if we focus on full-GPU processing, we can interop with GPU APIs, WebGPU is then our target and thanks to Chai's PR we have CommandEncoder for WebGPU, only remaining work is coordination with WebGPU WG for interoperability 14:29:25 #226 14:29:26 https://github.com/webmachinelearning/webnn/issues/226 -> Issue 226 Integration with real-time video processing (dontcallmedom) cr 14:30:00 dom: I consider #226 a WebNN issue, I don't expect WebRTC WG to have an opinion whether we can close this issue 14:31:01 ... I feel confident we can close this issue when we have integration story between the involved Web APIs clarified 14:31:32 q+ 14:31:44 ack ningxin_hu 14:32:17 ningxin_hu: to me, the major part is WebGPU and WebCodec interop and requirements 14:32:41 ... currently we don't have WebGPU interop as a "cr" GH issue 14:33:14 anssik: WebGPU interop is tracked as part of wide review 14:33:35 ningxin_hu: WebGPU group marked its #2500 issue "post-V1" 14:33:35 https://github.com/webmachinelearning/webnn/issues/2500 -> Issue 2500 [not found] 14:34:03 s|#2500|https://github.com/gpuweb/gpuweb/issues/2500 14:34:33 dom: how critical we think interop with WebGPU and WebCodec is, my perhaps biased perspective is it is critical 14:35:17 ... if we bring performant acceleration but introduce memory copies then some value is lost 14:36:26 ... I think wide review from WebGPU is not quaranteed, may be are you using our API properly, maybe not looking at the integration arch 14:38:26 anssik: so it sounds like we should open specific issues for interop with webcodecs and webgpu 14:38:32 ... mark them as cr dependencies 14:38:40 ... and remove that label from #226 14:38:41 https://github.com/webmachinelearning/webnn/issues/226 -> Issue 226 Integration with real-time video processing (dontcallmedom) cr 14:39:45 works for me 14:41:04 confirm 14:41:09 RESOLUTION: In response to real-time video processing #226, Ningxin to open new issues for WebGPU and WebCodec WebNN interop issues and remove "cr" label from #226 once done 14:41:28 #210 14:41:29 https://github.com/webmachinelearning/webnn/issues/210 -> Issue 210 Add method steps to operations (anssiko) cr 14:42:00 #128 14:42:01 https://github.com/webmachinelearning/webnn/issues/128 -> Issue 128 WebNN should support int8 quantized models (wchao1115) cr 14:42:33 +1 to being in scope. 14:43:42 Topic: CommandBuffer usage clarification 14:43:58 #264 14:43:58 https://github.com/webmachinelearning/webnn/issues/264 -> Issue 264 CommandBuffer usage clarification: internal, external, both? (bbernhar) 14:44:03 -> 2022/06/02 minutes https://www.w3.org/2022/06/02-webmachinelearning-minutes.html#t02 14:44:22 anssik: our proposed path was to have a note section in the spec to describe interaction and resource sharing rules for implementers 14:44:42 ... if this is still our best understanding, I'd like to update the issue with this or create a PR with the changes and ask Bryan's review 14:45:13 RafaelCintron: haven't had time to chat with Bryan yet and Chai is on vacation 14:47:13 Topic: Proposed new features 14:47:20 Subtopic: Support asynchronous context creation 14:47:25 #272 14:47:25 https://github.com/webmachinelearning/webnn/issues/272 -> Issue 272 Support asynchronous context creation (huningxin) cr 14:48:05 ningxin_hu: this is related to #229 and #230 14:48:06 https://github.com/webmachinelearning/webnn/issues/230 -> Issue 230 [closed] Should WebNN support async APIs? (huningxin) cr 14:48:06 https://github.com/webmachinelearning/webnn/issues/229 -> Issue 229 [closed] Should restrict the sync APIs to only exist in Workers? (huningxin) cr 14:48:30 -> related discussion in Chromium CL review: https://chromium-review.googlesource.com/c/chromium/src/+/3684745/7/third_party/blink/renderer/modules/ml/ml.idl#16 14:48:30 https://github.com//3684745/7/third_party/blink/renderer/modules/ml/ml.idl/issues/16 -> Issue 16 [not found] 14:48:53 RRSAgent, draft minutes 14:48:53 I have made the request to generate https://www.w3.org/2022/06/16-webmachinelearning-minutes.html anssik 14:50:06 ningxin_hu: The proposal is to introduce an async version of createContext and restrict the sync version to be only used by worker thread. 14:50:15 ... The model loader API has an async version of createContext. It would be better to have a consistent naming for sync and async methods between the two WebML APIs. 14:52:31 anssik: WebGPU uses async post-fix 14:53:49 dom: I checked there's only 2 examples of sync postfix 14:54:02 ... across all Web APIs 14:54:18 ... so consistency story is async postfix 14:54:47 ... given sync version is a facility for "non-JS developers", used for bringing non-JS codebases via worker to the web 14:55:02 ... this suggests making async default and use sync postfix 14:55:31 +1 14:55:34 instances of sync postfix: FileReaderSync(), CSSStyleSheet.eplaceSync(text) 14:55:40 q+ 14:55:44 s/eplace/replace/ 14:55:46 ack RafaelCintron 14:56:12 RafaelCintron: I agree we should make this method async 14:57:51 +1 14:58:21 i'll craft a PR for review 14:58:24 Subtopic: Support coordinate transformation modes for Resample2d 14:58:30 #270 14:58:31 https://github.com/webmachinelearning/webnn/issues/270 -> Issue 270 Support coordinate transformation modes for Resample2d (Honry) 14:58:54 anssik: Wanming requests: "Can we support these coordinate transformation modes in WebNN? At least for asymmetric, half_pixel_centers and align_corners. Or maybe we could refer to DML by defining InputPixelOffsets and outputPixelOffsets options in Resample2d to implement various coordinate transformation modes." 14:59:40 ... please comment on GH 14:59:43 Subtopic: Define the data type of the padding, strides and dilations 14:59:53 #269 14:59:53 https://github.com/webmachinelearning/webnn/issues/269 -> Issue 269 Define the data type of the padding, strides and dilations of MLConv2dOptions as sequence (miaobin) 15:00:05 anssik: Bin explains: "The data type of padding, strides and dilations of MLConv2dOptions are defined as sequence now. When we implement the backends of WebNN, we found that the data type of these arguments were defined differently by each backend. In fact, there is no need to use negative value for these three arguments, unlike axis in MLSplitOptions. So should we use sequence to define these three 15:00:05 arguments?" 15:00:34 -> Web Platform Design Principles > Use numeric types appropriately https://www.w3.org/TR/design-principles/#numeric-types 15:01:08 ... please provide your feedback in this issue 15:01:09 q? 15:02:07 Topic: TPAC 2022 15:02:14 -> W3C TPAC 2022 https://www.w3.org/2022/09/TPAC/Overview.html 15:02:21 anssik: 12–16 September 2022, hybrid meeting 15:02:32 ... In-person hub: 15:02:39 ... Sheraton Vancouver Wall Centre 15:02:47 ... Vancouver British Columbia, Canada 15:03:01 Subtopic: About TPAC 15:03:09 anssik: "This event brings together W3C technical groups, the Advisory Board, the Technical Architecture Group and the Advisory Committee for exciting, coordinated work. The benefit of assembling the community for thought-provoking discussions is invaluable." 15:03:26 ... "This year, the meeting will be hybrid with the main in-person hub in Vancouver, Canada, and remote event attendees." 15:03:32 Subtopic: Tentative schedule 15:03:38 ... Tentive WebML WG/CG schedule: 15:03:46 ... Tue 13 Sep 2022 15:00-18:00 Vancouver / 22:00-01:00 UTC / 6:00-9:00 Shanghai (Wed 14 Sep) 15:04:11 ... Tentative ML Ethics workshop sessions: Mon 12 & Thu 15 Sep 2022 afternoon(?) Vancouver time 15:04:21 ... Media & WebRTC WG joint meeting participation possible, an opportunity to discuss WebNN and WebRTC & Media joint work 15:04:25 thanks for the PRC friendly slot 15:04:37 ... such as integration with real-time video processing #226 15:04:38 https://github.com/webmachinelearning/webnn/issues/226 -> Issue 226 Integration with real-time video processing (dontcallmedom) cr 15:04:45 ... Detailed agenda for all the groups available soon 15:04:53 ... questions about TPAC? 15:05:02 RRSAgent, draft minutes 15:05:02 I have made the request to generate https://www.w3.org/2022/06/16-webmachinelearning-minutes.html anssik 15:05:46 RRSAgent, draft minutes 15:05:46 I have made the request to generate https://www.w3.org/2022/06/16-webmachinelearning-minutes.html anssik 15:07:10 RRSAgent, draft minutes 15:07:10 I have made the request to generate https://www.w3.org/2022/06/16-webmachinelearning-minutes.html anssik 17:03:14 Zakim has left #webmachinelearning