Can i use convolutional neural networks to make money

Thanks Ankush. We covered this a little in the previous post. I would be really grateful if you could answer my curiosity. Restricted Boltzmann machines, for examples, create so-called reconstructions in this manner.

By using our site, you acknowledge that you have read and understand our Cookie PolicyPrivacy Policyand our Terms of Service. Stack Overflow for Teams is a private, secure spot for you and your coworkers to find and share information. I am training a convolutional neural network to classify an image into one of five classes Class 1 — Class 5. I have very few training images for Class 1 and so I performed some data augmentation by taking random usw and flipping the images to create more data. I have at least training images for Class 2 — 5. Now, my training set consists of images for each class and I train it using stochastic gradient descent.

Your Startup

Convolutional neural networks. Sounds like a weird combination of biology and math with a little CS sprinkled in, but these networks have been some of the most influential innovations in the field of computer vision. Ever since then, a host of companies have been using deep learning at the core of their services. Facebook uses neural nets for their automatic tagging algorithms, Google for their photo search, Amazon for their product recommendations, Pinterest for their home feed personalization, and Instagram for their search infrastructure. However, the classic, and arguably most popular, use case of these networks is for image processing. Image classification is the task of taking an input image and outputting a class a cat, dog, etc or a probability of classes that best describes the image. For humans, this task of recognition is one of the first skills we learn from the moment we are born and is one that comes naturally and effortlessly as adults.

Towards Data Science

By using our site, you acknowledge that you have read and understand our Cookie PolicyPrivacy Policyand our Terms of Service. Stack Overflow for Teams is a private, secure spot for you and your coworkers to find and share information. I am training a convolutional neural network to classify an image into one of five classes Class 1 — Class 5.

I have very few training images for Class 1 and so I performed some data augmentation by taking random crops and flipping the images to create more data.

I have at least training images for Class 2 — 5. Now, my training set consists of images for each class and I train it using stochastic gradient descent.

I don’t expect a very accurate network given the limitations of my training set and images are probably not enough as well — but I would not have expected it to be so skewed given that Class 2 — 5 had the same number of distinct training images. If I had trained my network on a much larger proportion of Class 4 images then this would not surprise me. I would have expected the network to predict at least SOME of the other classes correctly.

This can happen although not very common. I think you have not trained long. A CNN tries to get one class at a time correctly, which is generally the one with the maximum number of samples if you have not normalized the loss. This is because it gets maximum benefit from predicting that class correctly in the beginning. As it gets better and better with time, it no longer gets that benefit and then tries to predict other classes correctly.

You can weight your softmax loss based on the class frequencies or re-sample your dataset to get around this problem. I also see that your CNN is not deep enough, the filter sizes are not appropriate for the resolution which you have and the number of training samples are too.

In practice, it is recommended to use small filter sizes such as 3×3, and 5×5 at the maximum since these give lesser parameters that decrease training time with no difference in accuracy as compared to 15×15.

There have been researches about this see ImageNet competitions winners On one hand, one possible reason that your model is biased to one class is because they are not balanced.

What you can do is penalized the model to be more biased to the class with smaller instances. Learn. Convolutional neural network making skewed predictions Ask Question. Asked 3 years, 8 months ago. Active 3 years, 3 months ago. Viewed times. My testing set consists of: Class 1 — 8 images Class 2 — 83 images Class 3 — images Class 4 — images Class 5 — images My network correctly predicts: Class 1 — 0 images Class 2 — 0 images Class 3 — 0 images Class 4 — images Class 5 — 0 images I don’t expect a very accurate network given the limitations of my training set and images are probably not enough as well — but I would not have expected it to be so skewed given that Class 2 — 5 had the same number of distinct training images.

Any thoughts? EDIT: Types of images: Buildings Network architecture: Input image — x x 3 Convolutional layer — 15 x 15 filters, 16 filters Max 2×2 pooling layer Convolutional layer — 11 x 11 filters, 32 filters Max 2×2 pooling layer Convolutional layer — 7 x 7 filters, 64 filters Max 2×2 pooling layer Fully connected layer — outputs Softmax classifier layer — 5 outputs Cost function: Cross-entropy.

There’s a specific datascience stack exchange — maybe your question would be better there? You should describe the architecture of your convnet as well as the objects you are classifying. EliKorvigo I have added those in as. Thank you!

Please report: exact sizes of training sets. Training errors crucial. Training method used. On the side note — your network looks quite simple small given the size of the input. Bharat Bharat 1, 1 1 gold badge 12 12 silver badges 31 31 bronze badges. Thank you for your comments and the links! I have re-sampled my dataset to get more Can i use convolutional neural networks to make money 1 images.

But you were right — my network is not being trained for long enough so for now it is basically predicting the last class it sees! Could you help me understand how deep my network should approximately be and what filter sizes are appropriate? Renz Renz 2 2 silver badges 7 7 bronze badges. Sign up or log in Sign up using Google. Sign up using Facebook. Sign up using Email and Password. Post as a guest Name.

Email Required, but never shown. How to create micro-interactions with react-spring: Part 1. This week, StackOverflowKnows syntactic sugar, overfit or nah, and the…. Featured on Meta. Thank you, Robert Cartaino. Change in roles for Jon Ericson leaving SE. Has Stack Exchange rescinded moderator access to the featured tag on Meta? How do the moderator resignations affect me and the community?

Visit chat. Related Hot Network Questions. Question feed. Stack Overflow works best with JavaScript enabled.

In this way, a net tests which combination of input is significant as it tries to reduce error. They also tend to give greater time to ML methods than actual factor selection when in actual fact the latter is the most important. We touched on four areas that you could think about in order to make money from machine learning: your own business, from social data, finance and gambling and competitions. You can play around with the shape, size and transparency however you like to create your perfect training dataset. Interested in reinforcement learning?