[Boards: 3 / a / aco / adv / an / asp / b / bant / biz / c / can / cgl / ck / cm / co / cock / d / diy / e / fa / fap / fit / fitlit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mlpol / mo / mtv / mu / n / news / o / out / outsoc / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / spa / t / tg / toy / trash / trv / tv / u / v / vg / vint / vip / vp / vr / w / wg / wsg / wsr / x / y ] [Search | Free Show | Home]

Neural network debugging

This is a blue board which means that it's for everybody (Safe For Work content only). If you see any adult content, please report it.

Thread replies: 12
Thread images: 2

File: 1485286320042.jpg (64KB, 900x900px) Image search: [Google]
1485286320042.jpg
64KB, 900x900px
How the fuck do I debug a neural network and keep track of everything so I can ensure that everything is working as intended? Is this how skynet begins?
>>
>>60665961
Pretty much. https://www.technologyreview.com/s/604087/the-dark-secret-at-the-heart-of-ai/
>>
Create a neural network to do the debugging
>>
Write a neural network that debugs neural networks.
>>
>>60666083
this is the correct answer
>>
File: 1484312600011.png (5KB, 405x335px) Image search: [Google]
1484312600011.png
5KB, 405x335px
>>60666083
>>
>>60665992
So I basically wasted my time writing it myself instead of using a library and focusing on how to actually apply it ? It gets to what I want but I don't understand how it does it.
>>
>>60666166
Oh, so this is bait
>>
>>60666107
It's neural networks all the way down, anon.
>>
>>60665961
I'm also having this problem OP. But the problem is mine doesn't work. It runs, but it seems like it's just training on noise.
I think my solution will be to initialize it with the same values as a known good network (copypasted, probably) and see where it goes off. It should be deterministic, hopefully

It sounds like you're having problems with understanding how the neurons actually work? Yeah, good luck with that. Deep dreaming was originally a debugging technique, so if you can make that work, you might be able to understand a little.
>>
>>60666166
you should follow courses online (free) or at least read a book to understand, even if not perfectly, the ways in which neural nets and other systems work. It's kind of complex, but not that much either.
>>
>>60665992
How much of this is sensationalism and how much is real?
Thread posts: 12
Thread images: 2


[Boards: 3 / a / aco / adv / an / asp / b / bant / biz / c / can / cgl / ck / cm / co / cock / d / diy / e / fa / fap / fit / fitlit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mlpol / mo / mtv / mu / n / news / o / out / outsoc / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / spa / t / tg / toy / trash / trv / tv / u / v / vg / vint / vip / vp / vr / w / wg / wsg / wsr / x / y] [Search | Top | Home]

I'm aware that Imgur.com will stop allowing adult images since 15th of May. I'm taking actions to backup as much data as possible.
Read more on this topic here - https://archived.moe/talk/thread/1694/


If you need a post removed click on it's [Report] button and follow the instruction.
DMCA Content Takedown via dmca.com
All images are hosted on imgur.com.
If you like this website please support us by donating with Bitcoins at 16mKtbZiwW52BLkibtCr8jUg2KVUMTxVQ5
All trademarks and copyrights on this page are owned by their respective parties.
Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.
This is a 4chan archive - all of the content originated from that site.
This means that RandomArchive shows their content, archived.
If you need information for a Poster - contact them.