14:01:55 <RRSAgent> RRSAgent has joined #webmachinelearning
14:01:55 <RRSAgent> logging to https://www.w3.org/2020/06/11-webmachinelearning-irc
14:01:56 <chai> chai has joined #webmachinelearning
14:01:56 <Zakim> Zakim has joined #webmachinelearning
14:01:57 <anssik> RRSAgent, make logs public
14:01:59 <ningxin_hu> Present+ Ningxin_Hu
14:02:02 <anssik> Meeting: WebML CG Teleconference – 11 June 2020
14:02:06 <anssik> Chair: Anssi
14:02:10 <anssik> Agenda: https://github.com/webmachinelearning/meetings/blob/master/telcons/2020-06-11-agenda.md
14:02:15 <anssik> Scribe: Anssi
14:02:22 <anssik> scribeNick: anssik
14:02:24 <anssik> Present+ Anssi_Kostiainen
14:02:29 <anssik> Present+ Chai_Chaoweeraprasit
14:02:35 <anssik> Present+ Rafael_Cintron
14:02:39 <anssik> Present+ Ping_Yu
14:02:47 <anssik> Present+ Ganesan_Ramalingam
14:02:59 <anssik> RRSAgent, draft minutes v2
14:02:59 <RRSAgent> I have made the request to generate https://www.w3.org/2020/06/11-webmachinelearning-minutes.html anssik
14:03:07 <RafaelCintron> RafaelCintron has joined #webmachinelearning
14:03:54 <anssik> RRSAgent, draft minutes v2
14:03:54 <RRSAgent> I have made the request to generate https://www.w3.org/2020/06/11-webmachinelearning-minutes.html anssik
14:04:05 <anssik> RRSAgent, make logs public
14:04:19 <anssik> TOPIC: Announced: W3C Virtual Workshop on Web and Machine Learning
14:04:34 <Baul_Eun> Baul_Eun has joined #webmachinelearning
14:04:35 <anssik> -> https://www.w3.org/blog/news/archives/8575 Virtual workshop announcement
14:04:45 <anssik> -> https://www.w3.org/2020/06/machine-learning-workshop/speakers.html#submit Apply to speaker
14:05:10 <anssik> anssik: W3C announced the Web & Machine Learning Workshop now being organized as a virtual event in September 2020.
14:05:24 <anssik> anssik: The event will be organized as a combination of pre-recorded talks (to be submitted in July 2020) followed by a series of live sessions in September, organized around 4 main themes:
14:05:39 <anssik> ... Opportunities and Challenges of Browser Based Machine Learning
14:05:58 <anssik> ... Web Platform Foundations for Machine Learning
14:06:12 <anssik> ... Machine Learning Experiences on the Web: A Developer’s Perspective
14:06:28 <anssik> ... Machine Learning Experiences on the Web: A User’s Perspective
14:06:53 <anssik> anssik: The event is free and open to anyone with relevant perspectives on the topic to register for the event. For more information on the workshop, please see the workshop details and submission instructions.
14:07:05 <anssik> -> https://www.w3.org/2020/06/machine-learning-workshop/ Workshop details and submission instructions
14:07:10 <anssik> anssik: Deadline to submit a proposal for a talk is 3 July 2020 and registration is open through 14 August 2020.
14:07:25 <anssik> anssik: Questions?
14:08:50 <anssik> TOPIC: LeakyReLU and min ops
14:09:01 <anssik> -> https://github.com/webmachinelearning/webnn/pull/63 Add leakyRelu and min (PR #63)
14:09:19 <anssik> anssik: PR reviewed and merged, thanks!
14:10:07 <anssik> ... Ningxin reports: Per the first wave models, leakyRelu is required by TinyYOLOV2 model. leakyRelu can be implemented by element-wise add, mul, max and min where min is being added in this PR.
14:10:16 <anssik> ... Any comments?
14:10:42 <anssik> ningxin_hu: Thanks Chai and Rama for review.
14:11:32 <anssik> anssik: I'd like to note we should add a more complex example to the spec that makes use of these new ops we've added to the spec recently. We could have a simple "Hello World!" example and in addition a more advanced one within a reasonable LOC limit.
14:12:03 <anssik> ningxin_hu: I support the idea to add a more complex example to the spec
14:12:18 <anssik> anssik: I'll make a issue for that
14:13:21 <anssik> s/a issue/an issue
14:13:26 <anssik> TOPIC: 2D pooling and reduce ops
14:13:40 <anssik> -> https://github.com/webmachinelearning/webnn/pull/64 2D pooling and reduce ops (PR #64)
14:13:58 <anssik> anssik: This PR defines the 2D version of pooling and relevant reduce ops
14:14:17 <anssik> ... other changes to make convolution's strides, padding, and dilations optional params for convenience
14:14:29 <anssik> ... also this PR received adequate review and was merged.
14:14:41 <anssik> ... Chai anything you want to mention about this PR?
14:15:18 <anssik> Chai: I decided to work on these ops together, since they're connected, and tried to make them consistent in terms of params and name a few optional
14:15:49 <anssik> ... pooling related to reduction so in one chart defined all the ops supported by all the frameworks, 7 types of reductions
14:16:14 <anssik> ... some frameworks were inconsistent among themselves (?)
14:16:30 <anssik> ... if we need to add a new reduction in the future can append this newly added section to keep consistent
14:17:27 <anssik> ... GlobalMaxPool is a specialized version of MaxPool and therefore it needs not be defined explicitly
14:17:45 <anssik> Chai: thanks Rama and Ningxin for your extensive review
14:17:52 <anssik> anssik: thanks for the work!
14:17:58 <ningxin_hu> q+
14:18:03 <anssik> ack ningxin_hu
14:19:08 <anssik> ningxin_hu: Some design considerations, conventions introduced in this PR, grouped related ops into one section such as reduction and pooling, in the spirit of this PR I put together a PR for softmax, in that PR the group was element-wise binary, add  etc. in one section
14:19:28 <anssik> ... convenient way to avoid spec definition duplication
14:20:07 <anssik> ... also element-wise unary for softmax, it is a separate PR, just making a point following the conventions
14:21:01 <anssik> q?
14:21:27 <anssik> TOPIC: Grouped conv2d
14:21:31 <anssik> -> https://github.com/webmachinelearning/webnn/pull/65 Support grouped conv2d (PR #65)
14:21:51 <anssik> anssik: This PR adds grouped convolution support to the existing conv2d definition, or convolutions in parallel said differently
14:22:03 <anssik> ... as with the previous PRs, our active participants reviewed also this PR ahead the call and it was merged
14:22:08 <anssik> ... great work!
14:22:52 <anssik> as a related matter, I'd like us to keep https://webmachinelearning.github.io/webnn/#acknowledgements updated
14:23:21 <ningxin_hu> q+
14:23:31 <anssik> ack ningxin_hu
14:23:50 <anssik> ningxin_hu: I'd like to add Rama to ack section
14:23:51 <anssik> +1
14:24:10 <anssik> s/+1/anssik: +1 to add Rama
14:24:36 <anssik> ... Ningxin anything to bring for discussion regarding this PR?
14:26:16 <anssik> ningxin_hu: I would like to highlight Chai's suggestion we call out depthwise conv op as a variant of grouped conv2d used in MobileNet first-wave model, a very good use of this spec
14:26:57 <anssik> anssik: any further comments?
14:27:05 <anssik> TOPIC: Noise suppression
14:27:11 <anssik> -> https://webmachinelearning.github.io/webnn/#usecase-noise-suppression Noise suppression use cases
14:27:28 <anssik> -> https://github.com/webmachinelearning/webnn/issues/66 Evaluate noise suppression models for required ops (issue #66)
14:27:42 <anssik> anssik: anyone interested in investigating this issue? Belem perhaps?
14:27:49 <anssik> anssik: some proposed tasks include:
14:27:55 <anssik> ... - Evaluate RNNoise model for required ops, to be added to the first-wave models
14:28:13 <anssik> ... - Gauge PoC interest, possible collaboration with related groups e.g. WebRTC
14:28:30 <chai> q+
14:28:35 <anssik> ack chai
14:29:08 <anssik> Chai: This is timely, Google Meet launched noise cancelling as a feature, an important scenario
14:29:22 <anssik> ... if nobody takes up the work, happy to look into RNNoise
14:30:24 <anssik> -> https://jmvalin.ca/demo/rnnoise/#music_player RNNoise PoC
14:31:18 <anssik> q?
14:32:25 <anssik> TOPIC: Adjourn
14:32:36 <anssik> RRSAgent, draft minutes v2
14:32:36 <RRSAgent> I have made the request to generate https://www.w3.org/2020/06/11-webmachinelearning-minutes.html anssik