|
have you had anyone @ed by them try to respond to the palin/trump bots? anything good?
|
#
¿
Feb 14, 2016 02:22
|
|
|
|
| # ¿ Apr 27, 2024 19:01 |
|
|
finally, the bot that replaces @dril
|
|
#
¿
Feb 14, 2016 12:52
|
|
|
|
|
#
¿
Feb 14, 2016 23:58
|
|
|
just wake up your boyfriend dude
|
|
#
¿
Feb 20, 2016 22:39
|
|
|
Randel Candygram posted:i'm working on a recurrent neural network shitposter. it's slow as balls to train, though, and i don't have nearly enough data i taught a cnn to make good posts using yospos as a training set, but it just lurks 
|
|
#
¿
Feb 25, 2016 19:22
|
|
|
hey crusader have you considered another nn that trains on wyrbot's questions and responses? i.e., you feed it the questions and have it predict the winner (or even the percentage split) then compare with the actual results. i don't know if there's enough training data but it would be interesting to see if you can become at all accurate at guessing which answer will win in advance. assuming it worked you could even use it to adjust wyrbot's questions, for instance if you wanted to try and ask fewer one sided questions. could even do something similar with the number of votes each question gets to try and ask better/more popular questions even if you don't want to alter wyrbot (why mess with perfection) it'd be cool to see if a network can "understand" human preferences on weird wyr questions
|
|
#
¿
Oct 24, 2016 13:44
|
|
|
my girlfriend is Legos posted:yeah, the obvious ones are great and i don't understand why you'd want to get rid of them it was just an example, i'm not sure that there's actually a percentage breakdown that you'd want to target
|
|
#
¿
Oct 24, 2016 15:36
|
|
|
you could for sure use a GAN approach but you could probably even use a CNN - check out deepmind's PixelCNN and WaveNet architectures, seems like something analogous would work for text
|
|
#
¿
Oct 25, 2016 00:02
|
|
|
marketing companies are probably even now hiring ML researchers to design networks that generate the most optimally dank and viral memes of all time rumour has it that DennysNet has already become self aware
|
|
#
¿
Oct 25, 2016 00:13
|
|
|
loss.jpg function
|
|
#
¿
Oct 25, 2016 00:14
|
|
|
gently caress libertarian bot is literally indistinguishable from an actual libertarian. if you could get it a camera and a fedora you could run a successful youtube channel with it
|
|
#
¿
Nov 8, 2016 23:00
|
|
|
libertarian bot is way too fuckin real
|
|
#
¿
Dec 15, 2016 15:54
|
|
|
came here to post this, lol forever
|
|
#
¿
Dec 29, 2016 16:37
|
|
|
Captain Foo posted:i like to think of this as a brutal mic-drop iceburn it p much is  e: oh wait is that you in the convo? well lol anyweay
|
|
#
¿
Jan 9, 2017 20:43
|
|
|
big scary monsters posted:libertarian bot is way too fuckin real
|
|
#
¿
Jan 15, 2017 12:06
|
|
|
lmao it's picking out a number of lines verbatim though, might be worth expanding the corpus. i'd suggest adding cannibal corpse and the backstreet boys
|
|
#
¿
Jan 16, 2017 11:54
|
|
|
i was thinking about using rap as a source but im not sure how id feel about creating a bot that says nigga
|
|
#
¿
Jan 16, 2017 23:02
|
|
|
getting some templeos oracle vibes from this now
|
|
#
¿
Jan 17, 2017 00:16
|
|
|
she was looking kinda dumb with her finger and her thumb ripping apart severing flesh gouging eyes tearing limb from limb
|
|
#
¿
Jan 17, 2017 02:28
|
|
|
https://twitter.com/eripsabot/status/821835852749873152 it's already beautiful
|
|
#
¿
Jan 19, 2017 01:05
|
|
|
https://twitter.com/eripsabot/status/821832501517807617 lmao
|
|
#
¿
Jan 19, 2017 01:06
|
|
|
can i choose neither? 
|
|
#
¿
Jan 23, 2017 23:09
|
|
|
https://twitter.com/WYR_bot/status/823757700119068676 https://twitter.com/WYR_bot/status/823576482748493825 who let smoka into the training corpus??
|
|
#
¿
Jan 24, 2017 16:37
|
|
|
Synthbuttrange posted:
gently caress
|
|
#
¿
Feb 14, 2017 13:39
|
|
|
like does it imply that if you eat the mom all day from space you'll be really constipated at the same time?
|
|
#
¿
Mar 7, 2017 13:08
|
|
|
it's a brave new world of piss
|
|
#
¿
Mar 8, 2017 10:32
|
|
|
eripsabot telling some hard truths about ai researchers
|
|
#
¿
Mar 11, 2017 19:23
|
|
|
Silver Alicorn posted:https://twitter.com/WYR_bot/status/840970701020549121 this is the hardest one yet
|
|
#
¿
Mar 12, 2017 19:18
|
|
|
WrenP-Complete posted:i cant even vote on this one because i keep on going back and forth. it was extremely difficult but in the end i had to go for the golden lab
|
|
#
¿
Mar 12, 2017 20:16
|
|
|
this changes everything
|
|
#
¿
Mar 12, 2017 21:14
|
|
|
been thinking about this for like the last half hour and still going to have to go with the dog
|
|
#
¿
Mar 12, 2017 21:52
|
|
|
but does erpisabot prefer panda or retriever? answer the important questions please
|
|
#
¿
Mar 13, 2017 15:55
|
|
|
NNs making up new names is the greatest thing.
|
|
#
¿
Apr 5, 2017 11:24
|
|
|
i'm glad that Spork Demo was successfully sold and turned into Spork Race, and Death of the Robot sounds like a candidate for 2018 indie goty also im Ksing Gelmor: The Time Spanger
|
|
#
¿
Apr 6, 2017 12:24
|
|
|
Synthbuttrange posted:Hey baby, I’m swirked to gave ever to say it for drive. quote:You are so beautiful that you make me feel better to see you.
|
|
#
¿
Apr 8, 2017 14:09
|
|
|
new youtube channel idea: cooking with neural networks episode one: cheese salad in sauce sauce (with beefer a laliar)
|
|
#
¿
Apr 8, 2017 15:15
|
|
|
WrenP-Complete posted:Earlier today eripsabot named himself Ken. ty for respecting your bots right to self determination
|
|
#
¿
Apr 21, 2017 11:44
|
|
|
i'm not completely sold on this bot yet but it does have its moments https://twitter.com/howtodoesthing/status/859898142426357760
|
|
#
¿
May 4, 2017 00:57
|
|
|
it's too late, 3d face recognition is here and it works pretty well even with beards and glasses these days. probably not quite cheap enough for mass use just yet but it's getting there
|
|
#
¿
May 8, 2017 17:00
|
|
|
|
| # ¿ Apr 27, 2024 19:01 |
|
|
i wrote a proposal to use x-ray imaging for automated detection of dead birds and rodents in crops (they make a mess of the harvesters) but my then boss took a look and decided that he couldn't face the amount of paperwork he'd have to do even to test the idea, let alone take the device outside the lab short sighted fool, i'll get my robot controlled death ray funded sooner or later
|
|
#
¿
May 8, 2017 21:10
|
|