Quantum Neural Networks

Quantum neural networks study how quantum states, gates, and measurements can be trained to perform tasks such as classification, forecasting, graph learning, and scientific modeling. Most systems are hybrid: classical networks prepare or process data, while parameterized quantum circuits encode features or serve as trainable model layers; common designs include data re-uploading, quantum convolutional circuits, and quantum attention. Work also explores how encoding, entanglement, circuit architecture, and measurement choices affect expressivity and trainability, with tests ranging from standard image and time-series benchmarks to real quantum hardware. A central tension is whether compact circuits can offer reliable benefits as qubit counts grow, given noise, costly training, and barren plateaus; convincing evidence requires comparisons with strong, parameter-matched classical models.