Neural Network (NN) is a specific architecture of machine learning, that has shown great successes in fields such as computer vision, speech recognition. Such advances are largely due to the availability of large amounts of data, as well as the rapid increase of computational powers. A well-trained NN can 'think' like a real human, as sometimes outperform human at certain tasks, such as the success story of AlphaGo.
In essence, many of the NNs are trained to recognize some patterns that are embedded in the data. The example below shows a simple case, in which a Convolution Neural Network (CNN) is trained to identify a handwritten number. The CNN model is directly trained and deployed in your web browser, and you can monitor the training progress. At the bottom of the page, you can 'write' your own number (after letting the model train for a while), and test the prediction of the CNN model.
This model uses the open source library ConvNetJS.
In essence, many of the NNs are trained to recognize some patterns that are embedded in the data. The example below shows a simple case, in which a Convolution Neural Network (CNN) is trained to identify a handwritten number. The CNN model is directly trained and deployed in your web browser, and you can monitor the training progress. At the bottom of the page, you can 'write' your own number (after letting the model train for a while), and test the prediction of the CNN model.
This model uses the open source library ConvNetJS.
Iteration Steps: