12:52:24 <RRSAgent> RRSAgent has joined #webmachinelearning
12:52:25 <RRSAgent> logging to https://www.w3.org/2019/04/11-webmachinelearning-irc
12:52:35 <Zakim> Zakim has joined #webmachinelearning
12:52:42 <anssik> RRSAgent, make logs public
12:52:51 <anssik> Meeting: WebML CG Teleconference – 11 April 2019
12:53:06 <anssik> Chair: Anssi
12:53:13 <anssik> Scribe: Venkat
12:53:22 <anssik> Agenda: https://github.com/webmachinelearning/meetings/blob/master/telcons/2019-04-11-agenda.md
12:53:33 <anssik> scribeNick: Venkat
12:53:42 <anssik> Present+ Anssi_Kostiainen
12:53:49 <anssik> Present+ Venkat_Narasimham
12:54:28 <anssik> RRSAgent, draft minutes v2
12:54:28 <RRSAgent> I have made the request to generate https://www.w3.org/2019/04/11-webmachinelearning-minutes.html anssik
13:17:47 <Venkat> me/ anssik shall be scribing details in this irc window.
13:25:24 <tomoyuki> tomoyuki has joined #webmachinelearning
13:26:05 <anssik> Present+ Tomoyuki_Shimizu
13:26:35 <Daniel_Smilkov> Daniel_Smilkov has joined #webmachinelearning
13:26:42 <anssik> Present+ Daniel_Smilkov
13:27:41 <Nikhil> Nikhil has joined #webmachinelearning
13:27:48 <anssik> Present+ Nikhil_Thorat
13:28:53 <anssik> Present+ Chris_Heilmann
13:29:32 <Rama__Microsoft_> Rama__Microsoft_ has joined #webmachinelearning
13:29:37 <gregwhitworth> present+
13:30:10 <anssik> Present+ Greg_Whitworth
13:30:35 <chrisheilmann> chrisheilmann has joined #webmachinelearning
13:30:40 <belem> belem has joined #webmachinelearning
13:30:42 <anssik> RRSAgent, draft minutes v2
13:30:42 <RRSAgent> I have made the request to generate https://www.w3.org/2019/04/11-webmachinelearning-minutes.html anssik
13:30:56 <anssik> Present+ Belem_Zhang
13:31:08 <NingxinHu> NingxinHu has joined #webmachinelearning
13:31:38 <NingxinHu> Present+Ningxin_Hu
13:31:49 <anssik> Present+ Ningxin_Hu
13:32:07 <anssik> RRSAgent, draft minutes v2
13:32:07 <RRSAgent> I have made the request to generate https://www.w3.org/2019/04/11-webmachinelearning-minutes.html anssik
13:33:30 <Venkat> we have a new scribe, please be nice and talk clearly and slowly
13:33:34 <Rama__Microsoft_> Present+ G_Ramalingam
13:34:13 <Venkat> we can use more than 30 minutes today, we have three topic.
13:34:57 <Venkat> TOPIC: Foundation specification
13:35:33 <Venkat> anssik: agenda topic proposed by Greg from Microsoft
13:35:38 <anssik> 1) Discuss & review concrete API proposals (WebNN Chromium POC, JIT proposal, others)
13:36:26 <anssik> https://github.com/webmachinelearning/webnn/issues/11
13:36:39 <anssik> -> https://github.com/intel/webml-polyfill/blob/master/docs/api.md WebNN Chromium POC
13:36:44 <anssik> -> https://github.com/webmachinelearning/webnn/issues/11#issuecomment-474363912 JIT proposal
13:37:44 <Venkat> anssik: kickoff next stage api sketch.
13:38:35 <Venkat> gregwhitworth talking
13:39:10 <gregwhitworth> https://github.com/DanielMazurkiewicz/WebAI/blob/master/README.md
13:40:32 <Venkat> greg: speaking on webnn proposal, use cases.
13:42:05 <gregwhitworth> gregwhitworth: Microsoft reviewed WebNN, WebAI, jit, and our own proposal
13:42:56 <gregwhitworth> gregwhitworth: We ultimately decided based on the use cases identified we propose use the WebNN Intel POC example
13:43:34 <gregwhitworth> gregwhitworth: we would like to use that as the foundation spec that we can evolve to meet the need for L1
13:44:41 <gregwhitworth> gregwhitworth: I'm fine allowing folks to review it but I'd like to make progress on it
13:45:09 <gregwhitworth> Nikhil: I agree folks are talking passed each other but that's due to the goals not being known
13:45:33 <gregwhitworth> Nikhil: we do know that clients want to execute ML on the client
13:45:56 <anssik> [ https://webmachinelearning.github.io/charter/ attempts to explain the target audience ]
13:46:06 <gregwhitworth> Nikhil: the angle I was coming from, it would be cool if this was an authoring platform in the browser
13:46:30 <Venkat> Nikhil_Thorat: programming world moving towards eager based approach
13:46:39 <gregwhitworth> Nikhil: I do think jit is on the trend and is entirely eager based, it's not about performance
13:47:29 <anssik> PROPOSED RESOLUTION: start "v0 spec" as a graph builder and graph-executor API and in "v1" explore direct usage of native calls in the browser as outlined in the JIT-based approach for improved web developer ergonomics
13:49:07 <chrisheilmann> +1 on that. I'd love to have the JIT earlier than later, from what I had as feedback people like the idea of using pre-defined models on device.
13:50:03 <gregwhitworth> gregwhitworth: yes that proposed resolution is what we'd like
13:51:30 <Venkat> tomoyuki speaking
13:51:51 <anssik> tomoyuki: graph builder and graph-executor API as first step sounds good
13:53:05 <Daniel_Smilkov> +1 to proposed resolution
13:54:18 <NingxinHu> +1 to proposed resolution
13:54:28 <tomoyuki> +1 to proposed resolution
13:54:41 <Venkat> RESOLUTION: start "v0 spec" as a graph builder and graph-executor API and in "v1" explore direct usage of native calls in the browser as outlined in the JIT-based approach for improved web developer ergonomics
13:55:02 <anssik> 2) Kick off concrete API sketching at https://webmachinelearning.github.io/webnn/
13:55:43 <Nikhil> no objections to the graph-based proposal v0 spec. I wanted to make sure we understand the loss of usability in favor of the simpler implementation (exposing the graph to the user) and it sounds like everybody understands those constraints to happy to move forward. We can iterate on the JIT / eager-based API in v1! :)
13:56:54 <anssik> PROPOSED RESOLUTION: Kick off concrete API sketching at https://webmachinelearning.github.io/webnn/ based on WebNN Chromium POC
13:57:33 <gregwhitworth> Proposed Resolution: Evolve specification using https://github.com/intel/webml-polyfill/blob/master/docs/api.md as foundation specification
13:59:04 <Venkat> anssik: thin wrapper over android nn api.
14:00:28 <anssik> PROPOSED RESOLUTION: Evolve WebNN API specification using https://github.com/intel/webml-polyfill/blob/master/docs/api.md as foundation specification
14:00:40 <anssik> q?
14:01:44 <Daniel_Smilkov> Starting with the WebNN API and evolving it sounds good
14:01:55 <gregwhitworth> gregwhitworth: Nikhil  or Daniel_Smilkov  - issues with that resolution proposal?
14:02:05 <Daniel_Smilkov> No issues
14:02:14 <anssik> [hearing no objections from anyone]
14:02:20 <Nikhil> +1 to Daniel, no issues
14:03:19 <Venkat> anssik: wait a sec, I'll reconnect to WebEc
14:04:51 <Venkat> RESOLUTION: Evolve WebNN API specification using https://github.com/intel/webml-polyfill/blob/master/docs/api.md as foundation specification
14:05:06 <NingxinHu> WebNN POC API starts as a thin wrapper of NNAPI. However it has main interfaces for a graph-builder and graph-executor, including Model, Compilation and Execution interfaces. There are many operations listed in the POC API doc. For WebNN foundation API, we can just start a very basic set of operations.
14:05:29 <Venkat> TOPIC: F2F meetings
14:05:53 <anssik> ScribeNick: gregwhitworth
14:05:54 <gregwhitworth> ScribeNick: gregwhitworth
14:06:11 <gregwhitworth> anssik: thank you Venkat
14:06:25 <gregwhitworth> anssik: what we want to do at the TPAC meeting
14:06:33 <gregwhitworth> anssik: it is scheduled for Sep 16-20
14:06:36 <anssik> https://www.w3.org/2019/09/TPAC/
14:06:37 <gregwhitworth> anssik: it's in Japan
14:06:51 <gregwhitworth> anssik: I did some research here
14:07:12 <gregwhitworth> anssik: WebGPU group does not have f2f at TPAC, and web assembly they will meet on Thursday
14:07:31 <gregwhitworth> anssik: so I propose that we would meet f2f on Tuesday 17th in the afternoon
14:07:47 <gregwhitworth> anssik: 1:30PM - 3:30, 3:30 - 5PM
14:07:59 <gregwhitworth> anssik: my plan is to merge these two slots into 4 hours of face time
14:08:08 <gregwhitworth> anssik: that may be a lot to ask since we're a CG
14:08:42 <anssik> PROPOSED RESOLUTION: WebML CG will meet F2F at TPAC Tuesday 17 September 2019 - 13:30-15:30 or 15:30-17:30, with preference to merge the two slots for a 4 hour session.
14:08:47 <gregwhitworth> anssik: Friday is another slot, but folks will be tired, but more energy  on Tuesday
14:09:26 <gregwhitworth> anssik: describes what TPAC is
14:11:06 <gregwhitworth> RESOLOLVED: WebML CG will meet F2F at TPAC Tuesday 17 September 2019 - 13:30-15:30 or 15:30-17:30, with preference to merge the two slots for a 4 hour session.
14:11:32 <gregwhitworth> RESOLVED: WebML CG will meet F2F at TPAC Tuesday 17 September 2019 - 13:30-15:30 or 15:30-17:30, with preference to merge the two slots for a 4 hour session.
14:16:27 <gregwhitworth> anssik: silence I think means consent
14:16:39 <gregwhitworth> RESOLVED: Have bi-weekly telecons
14:17:04 <gregwhitworth> NingxinHu: I heard from Jebs at Google it was a little early
14:17:13 <gregwhitworth> NingxinHu: a half an hour earlier is fine with me
14:17:28 <gregwhitworth> NingxinHu: half an hour later is fine with me
14:17:43 <gregwhitworth> chrisheilmann: maybe we can alternate times
14:17:56 <gregwhitworth> anssik: yes we can do that possibly
14:18:05 <tomoyuki> +1 to ningxin’s opinion :-)
14:18:19 <gregwhitworth> anssik: I would like to have active contribs on the call
14:18:42 <NingxinHu> I heard from James at Google it was a little earlier
14:20:16 <gregwhitworth> Resolved: push the meeting 30 minutes later for PDT ease of waking up
14:21:00 <belem> belem has left #webmachinelearning
14:21:01 <anssik> RRSAgent, draft minutes v2
14:21:01 <RRSAgent> I have made the request to generate https://www.w3.org/2019/04/11-webmachinelearning-minutes.html anssik
14:21:07 <Rama__Microsoft_> Rama__Microsoft_ has left #webmachinelearning
14:28:05 <anssik> Scribe: Greg
14:28:07 <anssik> RRSAgent, draft minutes v2
14:28:07 <RRSAgent> I have made the request to generate https://www.w3.org/2019/04/11-webmachinelearning-minutes.html anssik
15:31:59 <myles> myles has joined #webmachinelearning
16:09:32 <Zakim> Zakim has left #webmachinelearning