Fair deep reinforcement learning is provided. A microstate of an environment and reaction of items in a plurality of microstates within the environment are observed after an agent performs an action in the environment. Semi-supervised training is utilized to determine bias weights corresponding to the action for the microstate of the environment and the reaction of the items in the plurality of microstates within the environment. The bias weights from the semi-supervised training are merged with non-bias weights using an artificial neural network. Over time, it is determined where bias is occurring in the semi-supervised training based on merging the bias weights with the non-bias weights in the artificial neural network. A deep reinforcement learning model that decreases reliance on the bias weights is generated based on determined bias to increase fairness.

The present disclosure relates to systems, computer-implemented methods, and non-transitory computer readable medium for automatically transferring makeup from a reference face image to a target face image using a neural network trained using semi-supervised learning. For example, the disclosed systems can receive, at a neural network, a target face image and a reference face image, where the target face image is selected by a user via a graphical user interface (GUI) and the reference face image has makeup. The systems transfer, by the neural network, the makeup from the reference face image to the target face image, where the neural network is trained to transfer the makeup from the reference face image to the target face image using semi-supervised learning. The systems output for display the makeup on the target face image.

Techniques are described that extend supervised machine-learning algorithms for use with semi-supervised training. Random labels are assigned to unlabeled training data, and the data is split into k partitions. During a label-training iteration, each of these k partitions is combined with the labeled training data, and the combination is used train a single instance of the machine-learning model. Each of these trained models are then used to predict labels for data points in the k−1 partitions of previously-unlabeled training data that were not used to train of the model. Thus, every data point in the previously-unlabeled training data obtains k−1 predicted labels. For each data point, these labels are aggregated to obtain a composite label prediction for the data point. After the labels are determined via one or more label-training iterations, a machine-learning model is trained on data with the resulting composite label predictions and on the labeled data set.