Abstract
In this paper, a graphene-based terahertz (THz) absorber is presented using neural networks. The proposed structure contains graphene, which supports plasmon resonance, and provides tunable properties at the THz regime. Thus, applying a bias voltage to the designed absorber results in various frequency responses in the THz frequency spectrum. In order to predict the structural geometry of the proposed absorber in a fast and accurate way, artificial intelligence (AI) is employed. AI enables the possibility of designing a tunable THz absorber (when there is no limitation on applied bias voltage) and under a specific bias voltage. Several simulations using electromagnetic software have been conducted to generate a dataset for training the neural network. The resultant weights are then applied to define the absorbers' structures.