point-wise MLP

Point-wise MLP (Multi-Layer Perceptron) is a type of neural network architecture commonly used in natural language processing tasks, such as text classification and sentiment analysis. It is also known as feedforward neural network.

In point-wise MLP, each input data point (such as a sentence or a document) is represented as a fixed-length vector of features. These input vectors are fed into the network, which consists of multiple layers of non-linear transformations, including fully connected layers with activation functions. The output layer usually consists of a single neuron that produces a real-valued score, representing the predicted class or sentiment score.

The main advantage of point-wise MLP is its ability to capture complex non-linear relationships between input features and output labels. It can also be trained efficiently using backpropagation algorithm with stochastic gradient descent optimization. However, it may suffer from overfitting if the training data is limited or noisy.