Abstract
Entanglement distribution is crucial for quantum communication and cryptography but is hindered by channel loss and decoherence. Noiseless linear amplification (NLA) is a probabilistic protocol that supports noiseless amplification without violating the no-cloning theorem, aiding in tasks like entanglement distillation and enhanced metrology. The probabilistic nature of NLA and other quantum repeater elements creates a significant resource overhead, which depends on the success rate of each individual probabilistic step. We experimentally demonstrate a technique to increase NLA success probability while maintaining the amplification gain and the fidelity of the amplified state with a maximally entangled state. We show that symmetrically distributing loss before and after the amplification provides an improved scaling of the success rates of NLA. Our results are a critical step towards scalable quantum repeaters and enable efficient long-distance quantum communication