Figure 1
From: Neural network based successor representations to form cognitive maps of space and language
A scheme of the used Neural Network for the supervised and reinforcement learning task. The network receives the one hot encoded starting state of the environment as input. It uses one hidden layer with a ReLU activation function and the output is a softmax layer. In the case of the spatial exploration and the linguistic task, the output of the network is the one-hot encoded successor state of the corresponding input. In case of the spatial navigation task, which uses reinforcement learning, the output of the network is the action depending on the input state.
