|
Some more details about what this does step by step might help others to help you.
Nice code though.
Thanks.
|
|
|
|
|
Tutto è cominciato così:
1) ho dato in input al neurone javascript due numeri, due pesi e un bais.
2 ) il neurone ha calcolato un risultato che divergeva da quello da me previsto (target).
3) il neurone ha assottigliato, man mano, la divergenza fino ad arrivare a zero.
4 ) ora se rimetto al lavoro il neurone, mi da il risultato corretto.
5) questa operazione l'ho ripetuta per una serie di input con relativi target.
6) ho fatto la media di tutti i pesi e bais calcolati e mi ritrovo una retta.
Per ora sono fermo qui, ma il neurone non ha imparato nulla.
Qualche proposta ?
Riccardo
Translation:
It all started like this:
1) I gave input to the javascript neuron two numbers, two weights and a bais.
2) the neuron calculated a result that diverged from the one I predicted (target).
3) the neuron has thinned, gradually, the divergence until it reaches zero.
4) now if I put the neuron back to work, it gives me the correct result.
5) I repeated this operation for a series of inputs with relative targets.
6) I averaged all the weights and bais calculated and I find myself a straight line.
For now I'm stuck here, but the neuron hasn't learned anything.
Any proposals ?
Richard
|
|
|
|
|
Could A* be regarded as a Dijkstra algorithm that has the visited node chain spreading in all directions from the start point?
|
|
|
|
|
Of course, what's wrong with you?
The difficult we do right away...
...the impossible takes slightly longer.
modified 6-Nov-21 16:25pm.
|
|
|
|
|
|
No, it's the other way around. Dijkstra could be regarded as a special case of A*, with a heuristic of zero, resulting in a "circular" node visiting pattern (when possible), thanks to the heuristic of zero not "preferring" any particular direction. Some pedants may say "it's actually UCS that is a special case of A*, not Dijkstra", but UCS is a fancy name for how Dijkstra is normally implemented in practice. The original version of the algorithm that Edsger W. Dijkstra wrote down is rarely, if ever, used in practice, because it always touches all nodes, not just nodes that are dynamically encountered during the search.
|
|
|
|
|
Thanks for setting me on the track.
|
|
|
|
|
>>The original version of the algorithm that Edsger W. Dijkstra wrote down is rarely, if ever, used in practice, because it always touches all nodes
Is processing all nodes making the algorithm slow when you deal with large maps?
modified 8-Nov-21 10:34am.
|
|
|
|
|
I like A*.
I've build pathfinding with it for an AR HoloLens application.
Easy to build, easy to scale, easy to tweak.
It's also kinda crappy.
Just look at the AI in Red Alert 2: it's serviceable but never amazing.
|
|
|
|
|
I want to ask you in bioinformatics by machine learning. How do I extract features in protein primary sequence by Python?
|
|
|
|
|
What an interesting question! A have looking for an answer for the ast mounth 
|
|
|
|
|
VASCAR, THIS IS ALL ABOUT PRODUCT RIGHT ? WHAT IS THE MAIN CONCEPT USING AI ? ANSWER ME THIS BY STEP BY STEP.
|
|
|
|
|
If you want to ask a "generic" question, then start here: Ask a question[^]
But do yourself a favour, and learn how to ask a question so that it can be given a useful answer: Asking questions is a skill[^] and think about what you need to know, and what you need to tell us in order to get help.
And DON'T SHOUT. Using all capitals is considered shouting on the internet, and rude (using all lower case is considered childish). Use proper capitalization if you want to be taken seriously.
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
"Common sense is so rare these days, it should be classified as a super power" - Random T-shirt
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
|
|
Please don't SHOUT.
Give solutions to what? You need to give some more detail to your question, including explaining what you have already tried (so we don't just repeat stuff) and exactly where your problem is
|
|
|
|
|
|
Dear members,i would like to ask if is possible to train 2 different neural networks models(eg 2 cnn) and after doing the classification which is 2 outputs for each ,to retrain their weights from the classification layer in order make the final classification.Do you think it is a reasonable approach?any guideance to do this? Thanks
|
|
|
|
|
Please don't repost if your comment does not appear immediately: both of these went to moderation and required a human being to review them for publication. In order to prevent you being kicked off as a spammer, both had to be accepted, and then I have to clean up the spares. Have a little patience, please!
I've deleted the other version.
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
"Common sense is so rare these days, it should be classified as a super power" - Random T-shirt
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
where i can delete post
modified 14-Jun-21 13:31pm.
|
|
|
|
|
Sorry this is the wrong place. These forums are for discussing issues and problems around code that you have written. If you want people to write code for you then try rentacoder.com.
|
|
|
|
|
"Sorry this is the wrong place"it is your point of view
No need to regret.my post is to discussing, If you haven't seen this
I advise you to read the post again.
|
|
|
|
|
No it is not my point of view, it is the rules of the forum. But feel free to waste your time waiting for someone to do your work for you..
|
|
|
|
|
Member 15244735 wrote: after you finish that, let me know so i can share with you the other steps.
You mean, after we do your work you'll let us know what other work you want us to do for you.
 
|
|
|
|
|
"after we do" , Who do you mean by we? "your work" it is your point of view
No need to regret.my post is to discussing, If you haven't seen this
I advise you to read the post again
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
