A variant of self-supervised learning that is particularly useful when all of the following conditions are true:1
Self-training works by iterating over the following two steps until the model stops improving:
Use supervised learning to train a model on the labeled examples.
Use the model created in Step 1 to generate predictions (labels) on the unlabeled examples, moving those in which there is high confidence into the labeled examples with the predicted label.
Notice that each iteration of Step 2 adds more labeled examples for Step 1 to train on.
목차 목차 관련 글 Semi-supervised learning 이 글을 인용: “Self-training is one technique for semi-supervised learning.” ML crash course - Linear regression “The goal of training a model is to minimize the loss, reducing it to its lowest possible value.” Introduction to Machine Learning 관련 절: Definition of [[Machine learning|ML]] Example (machine learning) “Labeled examples are used during training.” A simple way to prevent neural networks from overfitting “Deep neural nets with a large number of parameters are very powerful machine learning systems.” Epoch (machine learning) “An epoch represents N/batch size training iterations, where N is the total number of examples.” Generalization (machine learning) “You train a model on the examples in the training set.” 모델 (기계학습) “In supervised machine learning, a model takes an example as input and infers a prediction as output.” Machine Learning Systems “Machine Learning Systems provides a systematic framework for understanding and engineering machine learning (ML…” TabNet: Attentive Interpretable Tabular Learning “Finally, for the first time to our knowledge, we demonstrate self-supervised learning for tabular data, signifi…” Iteration (machine learning) “When training a neural network, a single iteration involves the following two passes:” Clipping (machine learning) “Outliers can damage models, sometimes causing weights to overflow during training.” Downsampling (machine learning) “For example, before training an image recognition model, downsampling high-resolution images to a lower-resolut…” Training loss “A metric representing a model's loss during a particular training iteration.” Minority class (machine learning) “A training set with a million examples sounds impressive.” 배치 크기 “For instance, if the batch size is 100, then the model processes 100 examples per iteration.” Precision (machine learning) 관련 절: When (not) to use 레이블 “In supervised machine learning, the "answer" or "result" portion of an example.” 기계학습 관련 절: Generalization and OOD Test set “For instance, a single example shouldn't belong to both the training set and the test set.”