Modules use Machine Learning algorithms to learn to perform desired tasks. These algorithms can "learn" to make associations with those tasks based on a given set of examples or samples. With more practice in analyzing new samples, predicting results, and making corrections, they can eventually be trained to be more accurate.
MonkeyLearn provides users an easy way to train a module by providing samples for training, detecting points of confusion, and guiding users to where they can make corrections so they can retrain the module with new improvements applied.
For more on this topic, with examples, please see our Gentle Guide to Machine Learning blog post.