 Quantum neural networks, QNNs, offer a quantum advantage over classical artificial neural networks, ANS. They can process information faster due to their ability to exploit quantum effects like superposition and entanglement. However, the role of non-linearity in QNNs has not been fully explored. In this study, researchers investigated how cur non-linearity affects the performance of a bisonic QNN when simulating an XOR gate and generating Scrodinger cat states. They found that cur non-linearity reduces the effect of noise or losses, which are especially important in the quantum setting. Additionally, they noted that non-linear mapping could be achieved through a non-linear input-output encoding instead of a physical non-linearity, such as cur non-linearity. This article was authored by Hua Wenxiu, Tan Junkri's Nanda, Weichi Bao, and others.