100 votes

Crack CAPTCHA using deep learning

Here is one example using a convolutional neural network (CNN) to crack the CAPTCHA. We will use a CAPTCHA library to generate sample CAPTCHA images and then train a neural network to decode these ...
xslittlegrass's user avatar
59 votes

Object detection and localization using neural network

Introduction An object detection problem can be approached as either a classification problem or a regression problem. As a classification problem, the image is divided into small patches, each of ...
xslittlegrass's user avatar
41 votes
Accepted

Mathematica command that allows it to read my intentions

Indeed, this functionality still exists, but it has been moved into its own package. Load the package: Needs["aBetterProgrammer`"] You will have access to such ...
37 votes

Q&A Mathematica v.11 Neural Networks: A comprehensive look at Layers, Net Functions, and pioneering into this [[experimental]] code

Thank you for your summary. I would like to clarify and correct a few of your points. however, Mathematica - being proprietary - does not make it clear as to which algorithms they choose to use to ...
Sebastian's user avatar
  • 3,479
37 votes
Accepted

RNN in Mathematica?

This is actually a useful question, because it gives an opportunity for us to communicate with the community about what is coming, so we can get feedback. So here is some preliminary information about ...
Taliesin Beynon's user avatar
31 votes
Accepted

How to perform document classification (i.e. extracting topics from text)?

What you're describing is Topic modeling. Your link describes Latent Dirichlet Allocation (LDA), which is a popular model. You mention that you would like to use an expectation-maximization algorithm ...
V.E.'s user avatar
  • 1,700
30 votes
Accepted

Generative Adversarial Network

I wrestled with this for a while and got some kind of results, but nowhere near the great performance for which GANs are famous. Ultimately, they're absurdly sensitive to hyperparameters and ...
Michael Curry's user avatar
27 votes

How to export an MXNet?

It seems the model file in MXNet (checkpoint) is defined by two files: a ".json" file and a ".params" file. The json file contains the definition of the network, and the params file contains the ...
xslittlegrass's user avatar
26 votes
Accepted

Generalized Backpropagation for Neural Networks (e.g. DeepDream)

Sebastian mentioned in his answer that deepdream can be possible using NetDerivative. Here are my attempts following his outlines. Instead of using the inception ...
xslittlegrass's user avatar
24 votes

Mathematica command that allows it to read my intentions

Since version 11 most commands finally support the Interpretation option: Interpretation -> "Literal" being the classical (...
24 votes

Mathematica command that allows it to read my intentions

I tend to use a pattern matching approach: myCode/.{x_?BugQ:>BugStrip[x],x_?TypoQ:>Detypo[x],x_?WrongSignQ:>-x,x_?OffBy2PiQ:>x*2\[Pi]} With the usual ...
21 votes
Accepted

How to monitor the process of Neural Network Learning

You can pass the undocumented option "ShowTrainingProgress" -> False to turn off the training progress blob thing. As for monitoring the progress yourself, ...
Taliesin Beynon's user avatar
21 votes

Generalized Backpropagation for Neural Networks (e.g. DeepDream)

If you have an inception model, its mostly possible using hidden functionality (but without GPU training). The steps would look like this: Cut the inception model at some level using ...
Sebastian's user avatar
  • 3,479
21 votes

Mathematica command that allows it to read my intentions

This function was deprecated in V4.2, being succeeded by CellularAutomaton. Since your answer is hidden somewhere in rules like 110, why reinvent the wheel with <...
20 votes

How can I determine the importance of variables from Classify?

General remarks I am not aware of built-in functions for retrieval of the variable importance from/with Classify and related functions. I implemented (and ...
Anton Antonov's user avatar
20 votes
Accepted

Using a Convolutional Neural Network for time series classification

Here's the code. What I mentioned in the Q&A session is using ReshapeLayer to turn the input vector into a 1-channel, flat tensor that ConvolutionLayer can operate on, not to actually use images, ...
Taliesin Beynon's user avatar
19 votes

Q&A Mathematica v.11 Neural Networks: A comprehensive look at Layers, Net Functions, and pioneering into this [[experimental]] code

Layers BatchNormalizationLayer There are several layers introduced in v.11 that can not be used uninitialized, this is one of them. Input must be either a rank 1 or rank 3 tensor. To be honest, I do ...
SumNeuron's user avatar
  • 5,372
19 votes
Accepted

Recurrent neural network in 11.1 explicit examples?

Here is a simple example that may help you get started. In this example, we are going to a predict a simple time series of a sinusoid wave. ...
xslittlegrass's user avatar
18 votes
Accepted

How do I configure the input and output layers of a neural network?

I think it would be simpler to construct the simple neural network rather than setting the properties in Classify. Here is an example using a three-layer net to ...
xslittlegrass's user avatar
18 votes
Accepted

Importing a grid of numbers from an image (sudoku like)

1 - Summary of a simple solution In this particular DIGIT case there is a very simple solution based on neural nets (NNs)trained on MNIST Data. It is just a few lines of code: ...
Vitaliy Kaurov's user avatar
17 votes

What are the Wolfram Language's relative strengths for machine learning?

I actually failed miserably in a Kaggle contest using Mathematica Enterprise. I tested every single variation of Classify and Predict and even combinations of the two. I also tested Microsoft ML ...
ArtificialBrilliance's user avatar
17 votes

Generative Adversarial Network

Yes it is possible. You can do alternating training manually by literally following the algorithm, so that you have a Do loop whose body contains two calls to NetTrain, but that suffers from overhead ...
Taliesin Beynon's user avatar
17 votes
Accepted

How to use Mathematica to train a network Using out of core classification?

Okay here's how you do out-of-core training with HDF5: ...
Taliesin Beynon's user avatar
17 votes

Can we implement a Neural Network interactively?

Update: So after adding the missing features I decided to give your actual problem a go. This is what I have: You'll note a little "Add Layer" ActionMenu ...
b3m2a1's user avatar
  • 46.6k
16 votes

Noise filtering in images with PCA

Start data First let us get some images. I am going to use the MNIST dataset for clarity. (And because I experimented with similar data some time ago.) ...
Anton Antonov's user avatar
16 votes

How to make a custom NN layer in Mathematica?

Supporting custom layers is on our to-do list, and should be ready for either 11.2 or 11.3. For interest: what applications do you want custom layers for? And how performant do you need your custom ...
Sebastian's user avatar
  • 3,479
16 votes

How to use Mathematica to train a network Using out of core classification?

There are two parts to your question: 1. How to use out-of-core classification and 2. Why is the result bad. For the first part, you can use a generator to solve the problem. And for the second part, ...
xslittlegrass's user avatar
16 votes

Recurrent neural network in 11.1 explicit examples?

Taking inspiration from the answer by xslittlegrass, I came up with the following solution. Recall the sample data from this question. We have an observable obs we ...
Kagaratsch's user avatar
  • 11.9k
16 votes
Accepted

Deep learning outperforms in symbolic integration and ODE?

Not everything is as peachy as claimed in that paper due to limitations of the problems deep learning can approach in their framework. Wolfram Language (WL) addresses much wider range of problems. ...
Vitaliy Kaurov's user avatar
15 votes

How to use Mathematica to train a network Using out of core classification?

@xslittlegrass's answer is perfect, but I want to give a heads up that we will ship a way to stream training data to NetTrain from an ".h5" file that can be arbitrarily big (e.g. hundreds of gigabytes)...
Taliesin Beynon's user avatar

Only top scored, non community-wiki answers of a minimum length are eligible