Reversible Quantum Memory Compression: Hardware Validation of Ancilla Recycling via Uncomputation on NISQ Architectures

This paper investigates the viability of reversible quantum uncomputation as a mechanism for quantum memory recycling and data compression. The initial hypothesis proposed a "use-and-throw" circuit where redundancy created via entanglement could be stripped away (compressed) and later restored (decompressed) with zero data loss. We first detail a critical theoretical constraint observed during classical simulation: statevector simulators scale exponentially (O(2n)), causing "compressed" quantum states to artificially inflate in classical memory rather than shrink. Consequently, we tested the protocol on real IBM Quantum hardware (ibm_kingston). Using authentic data extracted from a 10 MB file mapped to rotation angles, we executed a 20-qubit circuit. The hardware successfully reset 10 ancilla qubits to the |0⟩ state with a 90.3% success rate, proving quantum memory recycling is highly viable. However, executing the full forward-andbackward cycle yielded only a 21.4% exact restoration rate. This disparity highlights the fundamental challenge of NISQ computing: while logical gates are mathematically reversible, physical noise is non-unitary and irreversible. We conclude that reversible uncomputation is an excellent tool for hardware-level qubit management, but cannot function as a classical file-zipping mechanism