TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Using neural nets to recognize handwritten digits

121 pointsby ivoflipseover 11 years ago

10 comments

svantanaover 11 years ago
Very well written, and I applaud the effort. But personally I don't care for the "magical" aura that writers tend to give ANNs - to me, they are simply (non-linear) function approximators that have a nice fitting algorithm. They work well for some problems and poorly for others. Also, beware of over-fitting - ANNs tend to be parameter-heavy, although there are approaches to prune the connections.
评论 #6797130 未加载
评论 #6798114 未加载
ivoflipseover 11 years ago
If you liked his first chapter, consider supporting his IndieGogo campaign for the whole book (<a href="http://www.indiegogo.com/projects/neural-networks-and-deep-learning-book-project/" rel="nofollow">http:&#x2F;&#x2F;www.indiegogo.com&#x2F;projects&#x2F;neural-networks-and-deep-l...</a>).
评论 #6795723 未加载
joe_the_userover 11 years ago
<i>&quot;The adder example demonstrates how a network of perceptrons can be used to simulate a circuit containing many NAND gates. And because NAND gates are universal for computation, it follows that perceptrons are also universal for computation.&quot;</i><p>I think this comment from the article needs caveats. Of course, a neural network would not qualify as Turing Complete just because it&#x27;s finite. Keep in mind also that neural network, lacking anything like counters, tape, or recursion, couldn&#x27;t approximate a Turing in the way that a finite Von Neuman architecture machine does. (A NN can represent any given function over a domain if it get large enough, kind of the universality of a finite automaton).<p>I know this a reference to this generation of NN having overcome an earlier problem of <i>not</i> being able to represent a NAND gate but still, it&#x27;s worthing keeping mind that an ordinary computer can simulate an NN with just a program but this doesn&#x27;t work vice-versa, so that NN&#x27;s in that sense are far from universal.
评论 #6798071 未加载
tbaover 11 years ago
This is a cool exercise! After completing it, I wanted to find out exactly what each NN hidden node represented. I trained a tiny (10 hidden node) NN on an OCR dataset and created a visualization here: <a href="https://rawgithub.com/tashmore/nn-visualizer/master/nn_visualizer.html" rel="nofollow">https:&#x2F;&#x2F;rawgithub.com&#x2F;tashmore&#x2F;nn-visualizer&#x2F;master&#x2F;nn_visua...</a> .<p>Can anyone figure out what each hidden node represents?<p>You can also select a node and press &quot;A&quot; (Gradient Ascent). This will change the input in a way that increases the selected node&#x27;s value. By selecting an output node and mashing &quot;A&quot;, you can run the NN in reverse, causing it to &quot;hallucinate&quot; a digit.
MechSkepover 11 years ago
What about convolutional neural nets? They weren&#x27;t mentioned, but that&#x27;s really what most of the deep learning approaches use...
评论 #6795633 未加载
cdurrover 11 years ago
Ahh, the MNIST database of handwritten digits. I never took an ML course and I was only able to achieve 87% recognition rate 6 years ago for a university software engineering project. I read others achieving 99.9% recognition rates with their ANNs so I wasn&#x27;t happy with my result. I tried to self-study about ANNs but I found most material to be either too simple or too complicated. I finally found some articles about ANNs (<a href="http://visualstudiomagazine.com/Articles/List/Neural-Network-Lab.aspx" rel="nofollow">http:&#x2F;&#x2F;visualstudiomagazine.com&#x2F;Articles&#x2F;List&#x2F;Neural-Network...</a>) with code samples in C#, so I&#x27;ll finally be looking into rewriting my old code to get a better result.
snarfyover 11 years ago
That reminds me of a video I saw about something called restricted boltzmann machines:<p><a href="http://www.youtube.com/watch?v=AyzOUbkUf3M&amp;t=24m0s" rel="nofollow">http:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=AyzOUbkUf3M&amp;t=24m0s</a>
评论 #6798228 未加载
Draco6slayerover 11 years ago
I&#x27;m not sure if I am making a mistake, but I couldn&#x27;t use the command listed in order to clone the repository. I&#x27;m using windows, with git installed, and I received the error: &quot;Permission Denied: publickey&quot;<p>I was able to get everything by looking you up on github and using the url of the repository.<p>Edit: Also, you might mention the repository earlier, because it&#x27;s rather large and I&#x27;ve had to break from the book while it downloads.
评论 #6798283 未加载
bcuccioliover 11 years ago
I did exactly this for a school project about a year ago:<p><a href="https://github.com/bcuccioli/neural-ocr" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;bcuccioli&#x2F;neural-ocr</a><p>There&#x27;s a paper in there that explains the design of the system and my results, which weren&#x27;t great, probably due to the small size of training data.
hcarvalhoalvesover 11 years ago
Isn&#x27;t that the same material covered on Andrew Ng&#x27;s Coursera course &quot;Machine Learning&quot;, down to the training data?
评论 #6797039 未加载
评论 #6797014 未加载
评论 #6796743 未加载
评论 #6796764 未加载