A blend of supervised machine learning and active learning where humans are involved in both training and testing stages of building an algorithm, combining the strengths of AI and human intelligence to enhance accuracy and effectiveness of the system.
In a deep learning project, HITL is used to train a model to classify images. The model is trained on a dataset of labeled images, but when it encounters an image that is difficult to classify, a human operator is involved in providing additional training data or correcting the model’s output.