Fig. 5
From: Hybrid quantum neural networks show strongly reduced need for free parameters in entity matching

Example quantum circuit of the HQNN model using four qubits. Before the “embed” barrier, the QAOAEmbedding layer encodes the input data into a state vector. In this example, the QAOAEmbedding layer is repeated twice. After the “embed” barrier, an entanglement layer is applied as an ansatz. The entanglement layer is PennyLane’s StronglyEntanglingLayers implementation.