Researchers ran IBM's Nighthawk quantum processor through two applied tests: a simplified model from particle physics and the filtering of malicious traffic. The experiments were noticed by Quantum Computing Report. In the first work, the team tried not just to "spin the qubits" but to compute a physical problem on the hardware — the interaction of a nucleon and an antinucleon in a simplified model of quantum chromodynamics, QCD2. The system was decomposed into a spin chain and run on Nighthawk. The resulting interaction potential showed the expected attraction and matched the results of classical checks — exact diagonalization and ideal simulation. The authors specifically emphasized that they extracted a useful signal from noisy data through structural error compensation. The second work is about cybersecurity and a more down-to-earth case: separating malicious DoS and DDoS traffic from normal traffic without "breaking" legitimate connections. The researchers took logs from a honeypot system and turned the task into a graph optimization, which they solved using the quantum approximate optimization algorithm, QAOA. The experiments used graphs of 16, 32, 66, and 110 events. The largest variant — 110 nodes and 181 edges — was run on three IBM backends from the IBM Quantum Network. According to the Quantum Computing Report summary, Nighthawk required the fewest two-qubit operations and produced the lowest compilation overhead, whereas the Heron-based processor showed a better target metric. The authors of both works do not claim quantum advantage. They present the results as an applied benchmark: how suitable such systems already are for tasks where both computational accuracy and noise resilience matter. As a reminder, in June IBM researchers described a new approach to searching for quantum error-correction codes using large language models. https://forklog.com/cryptorium/chto-takoe-kvantovye-vychisleniya-i-kvantovye-kompyutery

Source: ForkLog