Semi-supervised algorithms in action
We've discussed what semi-supervised learning is, why we want to engage in it, and what some of the general realities of employing semi-supervised algorithms are. We've gone about as far as we can with general descriptions. Over the next few pages, we'll move from this general understanding to develop an ability to use a semi-supervised application effectively.
Self-training
Self-training is the simplest semi-supervised learning method and can also be the fastest. Self-training algorithms see an application in multiple contexts, including NLP and computer vision; as we'll see, they can present both substantial value and significant risks.
The objective of self-training is to combine information from unlabeled cases with that of labeled cases to iteratively identify labels for the dataset's unlabeled examples. On each iteration, the labeled training set is enlarged until the entire dataset is labeled.
The self-training algorithm is...
