Researchers Find Quantum Data Loss Inequality Has Limits

Yu Cao and colleagues at the Shanghai Jiao Tong University have demonstrated that the tensorization property of the strong data processing inequality (SDPI) does not universally hold in the quantum realm, unlike its consistent behaviour with all classical divergences.

The team first established this property fails for certain quantum chi-square divergences and, crucially, does not hold for the quantum relative entropy. This indicates information loss within combined quantum systems does not consistently follow the same rules as in classical systems, specifically concerning how information degrades during processing. This discovery reveals a fundamental distinction between classical and quantum information theory and is vital for advancing the development of new quantum technologies. The strong data processing inequality (SDPI), a set of rules governing how much information can be lost when data is transmitted through a noisy channel, does not universally hold in the quantum realm. Specifically, the principle fails for certain quantum chi-square divergences, a way of measuring how different two quantum states are, and also for quantum relative entropy. The SDPI is a cornerstone of information theory, providing an upper bound on the rate at which information can be reliably transmitted through a noisy channel; exceeding this bound leads to inevitable information loss. Understanding the limits imposed by the SDPI is therefore critical for designing efficient and reliable communication protocols. Quantum information loss deviates from classical predictions in composite systems The strong data processing inequality (SDPI) constant, a measure of information loss, fails to tensorize for specific quantum divergences, achieving a value greater than 1/2 for combined systems. Previously, it was assumed to equal the maximum of the individual component rates. Tensorization, in this context, refers to the expectation that the information loss in a composite system (formed by combining multiple independent quantum systems) would be the maximum of the information loss in each individual component. The researchers found this is not the case; the combined system exhibits a greater information loss than predicted by classical tensorization rules. This challenges a long-held assumption that information loss in quantum systems behaves similarly to its classical counterpart, where tensorization consistently holds across a range of divergences. The value of exceeding 1/2 is significant as it represents a substantial deviation from the classical limit and highlights the fundamentally different behaviour of quantum information. Failures in tensorization were pinpointed for both quantum chi-square divergences and the quantum relative entropy, revealing a fundamental distinction in how information degrades during processing in quantum mechanics. Information theory predicts information loss through specific quantum channels, measured by the SDPI constant. Instances were identified where the SDPI constant for combined channels exceeded that of either individual channel, a result not observed in classical systems. This finding extends previous work indicating non-tensorization of contraction coefficients for quantum relative entropy, reinforcing a distinction between classical and quantum information processing. The quantum relative entropy, a measure of distinguishability between quantum states, is particularly important in quantum cryptography and quantum data analysis, making its non-tensorization a critical finding. Quantum data loss deviates from classical information theory predictions Reliable quantum communication and computing systems depend on understanding how information degrades as it passes through quantum channels. The research reveals that applying classical rules to these quantum systems is insufficient; the strong data processing inequality, a cornerstone of classical information theory describing information loss, doesn’t always behave as expected. These negative results do not diminish the importance of understanding how quantum information behaves differently from its classical counterpart. In fact, identifying these deviations is crucial for developing technologies that can harness the unique properties of quantum mechanics. Developing genuinely secure quantum communication networks and powerful quantum computers requires identifying where these classical rules break down. Data processing inequality is a fundamental property describing information loss through noisy channels, with a more refined description given by the strong data processing inequality. In classical information theory, tensorization of the strong data processing inequality holds for a family of f-divergences, but its quantum counterpart is less understood. Both quantum chi-square divergences and the quantum relative entropy, measures used to quantify differences between quantum states, demonstrate this tensorization property fails. The chi-square divergence, in particular, is useful for statistical inference and model comparison, while the relative entropy is central to quantifying the distance between probability distributions. The failure of tensorization for these divergences suggests that quantum information processing requires a fundamentally different theoretical framework than classical information processing. This work contributes to a growing body of evidence suggesting that quantum information theory is not simply a generalisation of classical information theory, but a distinct field with its own unique principles and limitations. The research demonstrated that the strong data processing inequality does not consistently tensorise for quantum chi-square divergences and the quantum relative entropy. This finding indicates that classical rules governing information loss through noisy channels do not fully apply to quantum systems. Identifying these differences between classical and quantum information processing is important for accurately describing how quantum information behaves. The authors showed that quantum information theory operates under principles distinct from its classical counterpart, requiring a separate theoretical framework. 👉 More information 🗞 Counter-examples for Tensorization Property of Strong Data Processing Inequality for Quantum Divergences ✍️ Yu Cao 🧠 ArXiv: https://arxiv.org/abs/2608.13204 Stay currentSee today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals. Tags: