Building a sample tensorflow.js application:

Introduction: TensorFlow is a machine learning framework for JavaScript applications. It helps us build models that can be directly used in the browser or in the node.js server. We use this framework for building an application that can predict request resolution time. 

Description: This JavaScript application uses data from a csv that has categorizations of requests and the resolution time. The attributes of each request include a category_id, a pseudo parameter attribute, and the execution time. This data is fabricated but it is also simplified to keep the application simple. 

As with any ML learning example, the data is split into 70% training set and 30% test set. There is no order to the data and the split is taken over a random set.  

The model chosen is a Sequential model. This model is appropriate for a plain stack of layers where each layer has exactly one input tensor and one output tensor.   Layers act in succession taking the output of one as the input of another. Thus, this model is suitable for one input and one output and where the layers are distinct and not sharing any input.  

Another model that can be used is one that is more generic and loads an acyclic graph.  A sequential model only uses a linear stack of layers. The input or output of the layers must be specified. A convolutional layer creates a convolution kernel which is a small matrix of weights. The kernel slides over the input layer and performs an element wise multiplication with the part of the input the kernel is on. The resulting scalar forms the element of a new kernel. This resulting kernel is called the convolution kernel. The graph model and sequential model can both be used with a tensor or a scalar. 

TensorFlow makes it easy to construct this model using an API keras Sequential. It can only present the hidden weight matrix after the model is executed. In this case, the model must be run before the weights are available.  The output of each layer can be printed using the summary() method.  

With the model and training/test sets defined, it is now as easy to evaluate the model and run the inference.  The model can also be saved and restored. It executed faster when there is GPU added to the computing. 

The features are available with the feature_extractor. It is evaluated on the training set using model.compile() and model.fit(). The model can then be called on a test input. Additionally, if a specific layer was to be evaluated, we can call just that layer on the test input. 

When the model is trained, it can be done in batches of predefined size. The number of passes of the entire training dataset called epochs can also be set upfront. It is helpful to visualize the training with the help of a high chart that updates the chart with the loss after each epoch 

When the model is tested, it predicts the resolution time for the given attributes of category_id and parameter attribute 

Conclusion: Tensorflow.js is becoming a standard for implementing machine learning models. Its usage is fairly simple but the choice of model and the preparation of data takes significantly more time than setting it up, evaluating, and using it.