Re: [silk] Steganography: This clever AI hid data from its creators to cheat at its appointed task

2019-01-03 Thread Charles Haynes
On Thu, 3 Jan 2019 at 14:40, Hariharan Rahul wrote: > I don't have much to say about the other discussion but the article really > got my teeth on edge, because of the tabloidy anthropomorphization of > something much simpler - "the loss function wasn't regularized properly > leading to

Re: [silk] it may not be well-done; is it becoming rare?

2019-01-03 Thread Charles Haynes
On Thu, 3 Jan 2019 at 19:08, Thaths wrote: > On Fri, Jan 4, 2019 at 9:47 AM Dave Long wrote: > > > > These days I think [email] is mostly used by us old fogies. > > Fair enough, but what, pray tell, do all those non-old-fogies use to > > convey thoughts that are too long for social media

Re: [silk] it may not be well-done; is it becoming rare?

2019-01-03 Thread Thaths
On Fri, Jan 4, 2019 at 9:47 AM Dave Long wrote: > > These days I think [email] is mostly used by us old fogies. > Fair enough, but what, pray tell, do all those non-old-fogies use to > convey thoughts that are too long for social media comments and too short > for blog posts? > Not being on

Re: [silk] it may not be well-done; is it becoming rare?

2019-01-03 Thread Heather Madrone
Dave Long wrote on 1/3/19 2:47 PM January 3, 2019: These days I think [email] is mostly used by us old fogies. Fair enough, but what, pray tell, do all those non-old-fogies use to convey thoughts that are too long for social media comments and too short for blog posts? -Dave They post

[silk] it may not be well-done; is it becoming rare?

2019-01-03 Thread Dave Long
> These days I think [email] is mostly used by us old fogies. Fair enough, but what, pray tell, do all those non-old-fogies use to convey thoughts that are too long for social media comments and too short for blog posts? -Dave

Re: [silk] Steganography: This clever AI hid data from its creators to cheat at its appointed task

2019-01-03 Thread Hariharan Rahul
I don't have much to say about the other discussion but the article really got my teeth on edge, because of the tabloidy anthropomorphization of something much simpler - "the loss function wasn't regularized properly leading to overfitting!" Rahul On Mon, Dec 31, 2018 at 9:09 PM Vinit Bhansali