Comments for Artificial Intelligence https://blogs.mathworks.com/deep-learning Johanna specializes in deep learning and computer vision. Her goal is to give insight into deep learning through code examples, developer Q&As, and tips and tricks using MATLAB. Tue, 06 Apr 2021 19:50:51 +0000 hourly 1 https://wordpress.org/?v=6.2.2 Comment on Deep Learning for Medical Imaging by SherlockSecure Bugcrowd https://blogs.mathworks.com/deep-learning/2019/07/24/deep-learning-for-medical-imaging/#comment-3024 Mon, 07 Oct 2019 06:08:16 +0000 https://blogs.mathworks.com/deep-learning/?p=2570#comment-3024 Thanks for sharing this.

]]>
Comment on What’s New in 19b: MATLAB by Steve Lord https://blogs.mathworks.com/deep-learning/2019/09/20/whats-new-in-19b-matlab/#comment-2982 Fri, 04 Oct 2019 20:10:02 +0000 https://blogs.mathworks.com/deep-learning/?p=2850#comment-2982 Related to what Heather wrote, if you’re interested in watching someone use Live Controls to build an interface for their code the most recent Pick of the Week blog post has a video where Sean de Wolski does just that.

https://blogs.mathworks.com/pick/2019/10/04/leaf-pile/

Creating and using the interface takes less than 100 seconds (the whole video is 1 minute 38 seconds long.)

]]>
Comment on Deep Learning with NGC by David Willingham https://blogs.mathworks.com/deep-learning/2019/05/03/deep-learning-with-ngc/#comment-2970 Fri, 04 Oct 2019 13:08:43 +0000 https://blogs.mathworks.com/deep-learning/?p=1943#comment-2970 UPDATE: R2019b container has now been released! It can be located here

]]>
Comment on Ensemble Learning by Johanna Pingel https://blogs.mathworks.com/deep-learning/2019/06/03/ensemble-learning/#comment-2882 Mon, 30 Sep 2019 13:27:51 +0000 https://blogs.mathworks.com/deep-learning/?p=2319#comment-2882 From Maria in response to calculating ensembling error, etc

% Model ensembling
% Average
predictionsMean = mean(predictionsAllModels,3);
getEnsembleError(predictionsMean, YValidation, labels);

% Majority voting
[m,~] = max(predictionsAllModels,[],2);
votes = sum(predictionsAllModels == repmat(m,1,size(predictionsAllModels,2),1),3);
getEnsembleError(votes, YValidation, labels);

% Weighted average (best model counts twice)
[~,bestModel] = min(validationError);
predictionsAllModels(:,:,end+1) = predictionsAllModels(:,:,bestModel);
predictionsMean = mean(predictionsAllModels,3);
getEnsembleError(predictionsMean, YValidation, labels)

Where:

function getEnsembleError(predictions, YValidation, labels)
[~,mi] = max(predictions,[],2);
predLabels = labels(mi);
validationErrorMean = mean(predLabels ~= YValidation);
disp("Ensemble validation error: " + validationErrorMean*100 + "%")
end

And predictionsAllModels was just the output (the probabilities obtained from ‘classify’) for all models:
predictionsAllModels(:,:,i) = probs;

]]>
Comment on Ensemble Learning by Hira https://blogs.mathworks.com/deep-learning/2019/06/03/ensemble-learning/#comment-2878 Fri, 27 Sep 2019 14:14:20 +0000 https://blogs.mathworks.com/deep-learning/?p=2319#comment-2878 Hellow, anyone helps me how can I calculate the ensembling error using average, weighted average and majority voting. what MatLab code is used? help me, please.

]]>
Comment on Image-to-Image Regression by William Bachman https://blogs.mathworks.com/deep-learning/2019/02/21/image-to-image-regression/#comment-2876 Wed, 25 Sep 2019 21:06:23 +0000 https://blogs.mathworks.com/deep-learning/?p=1280#comment-2876 I ran into that same problem Max, in the given code the regressionoutput isn’t actually named “regressionOutput, but instead is named “regressionLayer”. You can see this in the following section:

lgraph = lgraph.removeLayers(‘Softmax-Layer’);
lgraph = lgraph.removeLayers(‘Segmentation-Layer’);
lgraph = lgraph.addLayers(regressionLayer(‘name’,’regressionLayer’));
lgraph = lgraph.connectLayers(‘Final-ConvolutionLayer’,’regressionLayer’);

]]>
Comment on What’s New in 19b: MATLAB by Heather Gorr https://blogs.mathworks.com/deep-learning/2019/09/20/whats-new-in-19b-matlab/#comment-2870 Tue, 24 Sep 2019 18:22:04 +0000 https://blogs.mathworks.com/deep-learning/?p=2850#comment-2870 @Roman – Customization isn’t yet possible beyond selections and variable names in Live Editor Tasks. You could use Live Controls to build up something similar (or supplement an existing task) w/ dropdowns, check boxes, edit fields, etc.:
https://www.mathworks.com/help/matlab/matlab_prog/add-interactive-controls-to-a-live-script.html
I’ve also passed this request to the dev team, thanks for your question!

]]>
Comment on What’s New in 19b: MATLAB by Roman https://blogs.mathworks.com/deep-learning/2019/09/20/whats-new-in-19b-matlab/#comment-2868 Tue, 24 Sep 2019 16:31:05 +0000 https://blogs.mathworks.com/deep-learning/?p=2850#comment-2868 Live Editor Tasks: how can I setup a task myself? This is a question not adressed in the documentation. Thanks

]]>
Comment on MATLAB wins Hackathon by Johanna Pingel https://blogs.mathworks.com/deep-learning/2019/09/03/matlab-wins-hackathon/#comment-2794 Thu, 05 Sep 2019 12:20:28 +0000 https://blogs.mathworks.com/deep-learning/?p=2826#comment-2794 Thanks for keeping me honest Sebastian! The link is now fixed!

]]>
Comment on MATLAB wins Hackathon by Sebastian Bomberg https://blogs.mathworks.com/deep-learning/2019/09/03/matlab-wins-hackathon/#comment-2792 Thu, 05 Sep 2019 09:25:15 +0000 https://blogs.mathworks.com/deep-learning/?p=2826#comment-2792 Very impressive results those 97%. Congratulations to Paola and the team.

@Johanna: The link to the blog post on itility.nl is not working. The trailing period needs to be removed from the address, I believe.

]]>