News for hackers. Stuff that matters.

britcruise

britcruise on Hacker News

Re:AI Pets Are Here [video] on Tuesday August 04, 2026 @09:24PM (#49177495)
https://growbot.dev/start
Re:ChatGPT Learned to Reason [video] on Sunday December 01, 2024 @08:09AM (#42288269)
would love feedback on this
Re:How Neural Networks Learned to Talk – ChatGPT: A 30 Year History [v... on Sunday December 03, 2023 @09:39PM (#38513132)
Finally the algorithm found me
Re:I have question about information theoretic view of backprop on Friday October 11, 2019 @12:06PM (#21225676)
thanks for the comment.

Yes the performance boost in training is critical.

So what I'm saying is this performance boost is thanks to the switch to non-binary neuron (which allows a reversible operation to transmit a magnitude) - that's MOST important.

And separately ReLu are just better at this, because they are linear they don't have "vanishing edges" (which prevents the vanish/explode)

Separately I'm glad you brought up CNNs because CNN's are old, and go back to rosenblatt (1958), his perception had a first layer of local connections in it based on the findings in biological systems.

and of course that's because nature has found it's more efficient, and so the efficiency is huge.

but the point is CNN = fewer knobs to train.

and there are lots of simple ways to help reduce knobs - fix some connections - drop some connections

I have question about information theoretic view of backprop

Posted by britcruise on Friday October 11, 2019 @09:31AM
from the self-post dept.

I'm trying to explain the highest level reason why (modern) back propogation works better than previous methods (such as rosenblatts back propogation in 1958 - yes it's that old). without going into the calculus, i want to just look at the information side of things.I want to say something like the following:back in the 1960's we tried back propogation with binary neurons. And so when we tuned...

( Read More... | 2 comments | 2 points | on HN )
Re:I spent 16 months on this Bitcoin explainer. by Art of the Problem on Tuesday May 29, 2018 @09:20AM (#17178867)
It's because of trust - Bitcoin is a community not a thing.
Re:I spent 16 months on this Bitcoin explainer. by Art of the Problem on Tuesday May 29, 2018 @09:19AM (#17178863)
Thanks for the feedback! I actually had a different intro I chopped off - it's here: https://www.youtube.com/watch?v=TXB_j-_jPmU&t=1s

Older »