
33:09
list of ai and iot conferences is very doable - where LF DL can be highlighted.

33:32
O’Reilley AI conferences are important

33:41
LF conferences - of course - as well

34:14
For mission statement - some references that seem relevant - WSJ, “Models will run the world” - see https://www.wsj.com/articles/models-will-run-the-world-1534716720

35:56
Thanks Jim!

37:31
AAAI and NIPS are two important conference for researchers in AI and DL

38:08
Not easy to get a ticket for NIPS…

41:01
re: NIPS, joked with IBM colleagues that top rejected papers from NIPS would create a second A-level conference

42:44
yes, it is true for some top AI confs. AAAI has about 7000 submission this year

42:52
Candidate for LF DL: MITRE - see https://www.mitre.org/capabilities/advanced-technologies/symphony-automated-secure-platform

52:59
Thanks everyone, these points are helpful

56:27
Is ONNX in the cha\rt?

59:28
Agreed - on same layer as TF, PyTorch at this stage of diagram

01:01:07
The one thing to add to the chart is something about Memory LSTM, and some of the GANS stuff

01:01:43
Sorry I have to drop - Sept 27th I will be in ONS Europe as I am giving a talk there.

01:02:10
Technically hot area - is one model that does multiple NLP and Vision tasks

01:02:34
See these:

01:02:35
Richard Socher (Salesforce)Topic: From Narrow to Broad AI: An NLP AI model that can do multiple tasksWebsite: https://einstein.ai/research/the-natural-language-decathlonAlternative: Reza Zadeh (Matroid and Stanford) Image Recognition Tasks Paper https://arxiv.org/pdf/1804.08328.pdfData: http://taskonomy.stanford.eduVideo Image: https://twitter.com/Reza_Zadeh/status/1009189923432198144/video/1Alternative: Video: Lukasz Kaiser (Google, Tensor2Tensor)Video: One Model to Learn It All: https://www.youtube.com/watch?v=8FpdEmySsuc

01:02:59
^^^ for LF DL to be cutting edge