sno bunny

Tue Feb 18 18:42:01 2020
<773ab1f1> <https://machine-learning-and-security.github.io/papers/mlsec17_paper_51.pdf> &lt;— backdoored neural net.

Tue Feb 18 18:22:52 2020 <9e126bf3> <https://openai.com/blog/deep-double-descent/> — Deep Double Descent — We show that the double descent phenomenon occurs in CNNs, ResNets, and transformers: performance first improves, then gets worse, and then improves again with increasing model size, data size, or training time. This effect is often avoided through careful regularization. While this behavior […]

Tue Feb 18 17:57:25 2020 <8f79fcda> if you’re in a sales situation <8f79fcda> they don’t want to be told the answer, they want to feel smart and come to the conclusion on their own. <8f79fcda> how? <8f79fcda> find a champion of the capability – someone who sees the end of the rainbow <8f79fcda> hmmm… let […]

Tue Feb 18 16:34:14 2020 <8f79fcda> let me start by saying that i’m happy you called this chan machine learning not “AI” <8f79fcda> cuz AI is a farce <773ab1f1> So I had a quick convo with some guy on that exact topic on monday. <773ab1f1> In the biz, someone asking for AI is akin to […]

Tue Feb 18 16:26:35 2020
<d666283b> There are potential sets out there but there is no data like field data
<8f79fcda> That’s an interesting one…
<8f79fcda> I’m driving but I’ll give it some thought
<8f79fcda> In short mo…

Tue Feb 18 16:05:02 2020
<773ab1f1> <@U036M05JL> You seen any info on how companies are valuing their ML models on their books?

Tue Feb 18 15:09:12 2020
<773ab1f1> Yea but where would you get the training data?

Tue Feb 18 15:01:28 2020
<773ab1f1> Keep ur eyes peeled for a #MLtwitter list. I need to beef up my network

Tue Feb 18 14:54:34 2020
<773ab1f1> Then he stays!

Tue Feb 18 06:42:47 2020
<5547a085> Lol Trevor is cool, he rarely signs in to this slack

Tue Feb 18 06:05:34 2020
<773ab1f1> GUESS WHO JUST HAD A CONVO WITH MICROSOFTS ML/SECURITY TEAM LEAD WHO IS HIRING
<773ab1f1> hint: he has two thumbs

Tue Feb 18 05:49:09 2020
<773ab1f1> <https://machine-learning-and-security.github.io/papers/mlsec17_paper_51.pdf>

Tue Feb 18 05:37:56 2020
<d666283b> Using optical gas imaging cameras to detect leakage and send out alerts to operators if one occurs.
Ideally use computer vision to start training models to identify how many parts per million the leak is

Tue Feb 18 05:24:14 2020 <d666283b> On the down low, I’m talking to a potential oil and gas client that wants a computer vision solution for detecting hydrocarbon leakage <773ab1f1> Crap trump thought I invited you. I think it got trevorwt instead. Whoever he is <773ab1f1> How ya approachin the problem zorak? <773ab1f1> kick out […]

Tue Feb 18 05:05:24 2020 <773ab1f1> dude you would be amazed at how the image recongition works in ML <773ab1f1> First you gotta do a bit of a primer on image processing pre machine learning <773ab1f1> Some of the training methods to shorten training times start with like 4 pixels by 4 pixels and slowly […]