Stack Exchange Network

Stack Exchange network consists of 174 Q&A communities including Stack Overflow, the largest, most trusted online community for developers to learn, share their knowledge, and build their careers.

Visit Stack Exchange

Questions tagged [machine-learning]

Questions about machine learning functionality of Mathematica, for instance involving, but not restricted to, Predict, Classify, FindClusters and the various regression functions.

1
vote
1answer
213 views

Save trained NetChain between Mathematica sessions?

Let's say I have written a NetChain machine learning setup, trained the chain with NetTrain and obtained a model ready to ...
0
votes
1answer
86 views

generator didn't return a correct type in NetGraph with Multi-port

In this example,generator[<|"BatchSize" -> 4|>] get an association of lists, but NetTrain[net, generator] can't be used....
5
votes
1answer
816 views

How to make a switch in Neural Networks?

In this paper,they design a network like this. The left(green) is a normal network that mapping one feature to another feature,the right(red) side is an autoencoder(using RNN).But a switch is ...
3
votes
1answer
145 views

Is it possible to speed up Classify with PerformanceGoal->“Quality”

I want to reproduce a Kaggle's python version's result of MNIST. PCA+SVM with accuracy=0.98 on test set. By default in Classify, with SVM method, I get a result accuracy 0.9675 and now I get 0.975 ...
7
votes
1answer
265 views

Can we implement a Neural Network interactively?

Implementing a Neural Network manually need some code and need to make some connections between these components. Can we implementing a Neural Network interactively?If I want to make a net like ...
0
votes
0answers
63 views

How can I improve effect of Clustering of MNIST?

Limit us with Clustering technology, like FindClusters, ClusteringComponents,ClusterClassify,...
15
votes
5answers
2k views

Recurrent neural network in 11.1 explicit examples?

I heard that RNN was implemented in Mathematica as of 11.1. Trying to search online, I find some general information about neural networks in Mathematica, or a list of related functions. My trouble is ...
2
votes
0answers
236 views

Predict a lagging continuous observable based on leading continuous parameters?

Consider a time dependent observable obs[t] and a set of n time dependent parameters par[i,t]...
7
votes
1answer
397 views

What does the value of Predict's “L2Regularization” correspond to?

What does the value supplied to the Predict Method option "L2Regularization" correspond to? ...
4
votes
1answer
155 views

Why LogisticRegression in Classify gives worse accuracy than result without PerformanceGoal Setting

Tested@Mathematica 11.1 on Windows ModelA ...
3
votes
0answers
76 views

Transfer Learning: How to take certain layers from a pre-trained neural net, add it to a new net, and keep its variables fixed? [duplicate]

How can one copy specified layers from a pre-trained neural network and add them to another trained neural net? For instance, ...
4
votes
0answers
139 views

NetTrain with External GPU?

I'd like to use the new Neural Networks functionalities in v11 on my Macbook Pro, unfortunately, most macs do not have the supported Nvidia cards. However, there is the option for using an external E-...
3
votes
1answer
74 views

Is there a good explanation of the output of FeatureExtract anywhere?

I have been looking at the official documentation of FeatureExtract, and I'm having a bit of difficulty interpreting the output--or understanding the reasons for ...
4
votes
1answer
501 views

Is it possible to fix the weight of a certain layer during neural network training?

As suggested by the title, is it possible to fix some part of the neural network while training? Since Mathematica provides a way to extract part of a neural network and combine it with some layers ...
4
votes
1answer
89 views

Accuracy during training?

Is there any way to see a plot of the accuracy over the Validation set for each round instead of the loss graph, during a NetTrain session? There probably is and I'm just missing it...
8
votes
1answer
307 views

NetTrain automatic settings

I'm trying to compare Mathematica's NetTrain with other algorithms, and I used Automatic settings as they were working quite ...
8
votes
1answer
384 views

Batching in inference and support for multiple GPUs

I know that Mathematica supports GPU inference, and can be told which GPU to run models on by specifying TargetDevice->{"GPU",2} or such. My questions are as follows: 1) By using the above statement ...
2
votes
1answer
71 views

ImageIdentify category specification

There is an instance in the reference for ImageIdentify were a category is specified. Here it is: ...
3
votes
1answer
385 views

Joint Training and multiple losses in NetTrain

Jointly training a neural network to solve multiple tasks simultaneously sometimes results in elevated accuracy, and I'm looking for any known examples of this in Mathematica. Here's the specific ...
1
vote
1answer
124 views

Need advice for approach for multivariate regression/prediction using Deep Networks

I am trying to build predictive models for 2 outputs given approximately 10 input variables. I understand that I would first look at the simplest model, such as a multivariate linear regression, ...
3
votes
2answers
202 views

Regarding input to NetTrain function

Essentially I have used the SemanticImport[] function on a csv file with headers for each column. Now I understand that each of these headers is referred to 'key' in mathematica. * I have no problem ...
1
vote
1answer
218 views

Problem initializing training set on recurrent net

I would like to train a time-series data using GatedRecurrentLayer. The data consists of k scalar features mapping to a scalar number at every time step: i.e. ...
10
votes
3answers
621 views

Using Mathematica wlnet model in C++

Suppose if there is a trained net: SeedRandom[1234]; net = NetInitialize@NetChain[{5, 3}, "Input" -> 2]; It can predict image in mma: ...
1
vote
0answers
109 views

Reducing the size of the training data

My problem is that running Classify on my training data takes too long. First I get 50,000 images such as car, cat, dog, hat, train, etc. ...
10
votes
1answer
2k views

Markov Random Field

I look for an efficient way to implement a Pairwise Markov Random Field. I have implemented a homogeneous (i.e., unary potential is a similar function in all nodes, and pairwise potential is a ...
0
votes
0answers
164 views

Rectangular object boundary detection

I would like to find the boundaries of an object: I have looked that the YOLO algorithm but it looks like the results cut off parts of the object a lot. What kind of algorithm can I use to achieve ...
3
votes
1answer
137 views
0
votes
0answers
55 views

How can I speed up import images into memory and make training more efficient

My memory is 32GB, windows 10, cpu 7700k, GTX1080 50k images[400MB] cost me 30 minutes. In a CAPTCHA example, a simple net and in tunning hyper-parameters step with small values of MaxTrainingRounds, ...
1
vote
0answers
195 views

NetTrain always crashes in out of core training with NetEncoder Setting

See this official example example My enviroment is Windows 10 GTX1080 with Mathematica 11.1.1 It's Ok if I import all training images into memory. Even in the simple case...
3
votes
1answer
172 views

ClassifierMeasurements Indeterminate Examples

I like to use ClassifierMeasurements. It's nice and doesn't make me think too hard. One of the features I like a lot is the ...
3
votes
1answer
199 views

Specifying mean and kernel function in Gaussian Process

As part of the Improved Machine Learning in Mathematica 11, Wolfram advertises the Method "GaussianProcess" which can be used with Predict[] to obtain smooth Gaussian random fields. The screenshot ...
11
votes
2answers
482 views

How to combine Neural Network and Audio like Classify?

I want to combine Audio object and Neural Network ,first thing I remember is Classify ...
1
vote
1answer
114 views

How to get top-n accuracy?

Is there a way to get the top-3 or top-5 accuracies per class from a model using ClassifierMeasurements[]?
2
votes
0answers
73 views

Is there a way to monitor progress of Classify?

NetTrain provides great feedback while doing gradient decent or any other training method. Is there a similar way to monitor training details with a call to ...
8
votes
1answer
135 views

How to extract of sub-net of NetGraph correctly?

net = NetGraph[{3, 5, 8}, {1 -> 2 -> 3}, "Input" -> 2] Take[net, {1, 2}] But ...
5
votes
2answers
242 views

How to make similar objects have similar embedding weights?

In the read world, there are many many phones(aa ae ah ao ar aw ax ay b ch d dh ...). But for brief,I only use three phones(aa ah ae) to illustrate my question.And it only need training Data,no Test ...
3
votes
0answers
97 views

How can I monitor the process of neutral network training?

In the Linux command line,type wolfram then use NetTrain[net, mnist, Automatic, "LossEvolutionPlot"] can get the ...
3
votes
1answer
84 views

Similar data,same net,different level of MSE loss?

Using same net and similar data,but MSE loss is very different. This is the original data,and it can not learn very well. ...
11
votes
2answers
380 views

How to make Advanced Activations Layers in Mathematica?

I want to try Advanced Activations Layers in Mathematica,but not found. So I try to use ElementwiseLayer to implement it. ...
8
votes
1answer
188 views

How does mathematica compute loss with multi-losses?

I want to make a DNN that it can learn from text feature(1534-dims) and old audio features(256*4(four old frames)-dims) to predict audio feature(256-dims) at current frame. So it has two inputs,one ...
12
votes
2answers
393 views

Simple example of NetGraph running in MXNet?

There is an great example of exporting a NetGraph (that was trained in Mathematica) into the MXNet's format. However, I think there's a bug or something is wrong with the answer, because once ...
12
votes
1answer
296 views

Using PredictorMeasurements with a neural net?

PredictorMeasurements doesn't work with NetGraph, here's an example: ...
0
votes
0answers
99 views

Cross-entropy calculation

This is a really dumb question, but here goes: is there a way to compute cross-entropy of two probability distributions? There is CrossEntropyLayer[], so ...
19
votes
1answer
1k views

How to import a Tensorflow model?

I've trained a network in Tensorflow and have the checkpoint files, I'd like to if anyone has written a parser or importer to pull the evaluation graph and/or weights into Mathematica?
2
votes
0answers
91 views

Fresh install: There is no $DefaultLocalBase value set [closed]

Fresh install Mathematica 11.1.1.0 on Windows 10 The command $DefaultLocalBase returns $Failed and an error message ...
1
vote
1answer
351 views

stupid `NetTrain[]` question

Consider this sequence: ...
2
votes
0answers
65 views

Machine learning of “>” [closed]

Here is what should be a simple problem with the current state of technology: ...
3
votes
1answer
222 views

How to use Multi-Inputs in NetGraph with NetTrain?

I want to train a network,the first input is a Integer,the second input is 10-dims vector. Assume we have those data: ...
3
votes
0answers
75 views

ClassifierMeasurement gives reversed labels for precision and accuracy

According to the standard definitions of precision and recall, Precision = TruePositives/(TruePositives+FalsePositives) Recall = TruePositives/(TruePositives+FalseNegatives) In terms of confusion ...
2
votes
1answer
105 views

How to get low dimension representation of data based on some knowledge?

I want to map the 500-dims unit vectors to 128-dims and it know some knowledge(43-dims). So I write this code: ...