Two dice always land on the same number, no matter how far apart you toss them. In quantum physics, this behavior is called entanglement, and it's the key resource for future computers and secure communication. The trouble is, not all entanglement is easy to spot. Some forms are like invisible ink: the connection is there, but a standard test (called the "positive partial transpose criterion") shows a blank page.
To reveal such hidden entanglement, you need special tools — positive but not decomposable maps. In plain terms, these are mathematical maps that, when applied to one half of an entangled pair, can highlight the connection. Constructing such a map by hand is like picking a lock with a thousand pins. The authors turned this search into a task for differentiable semidefinite programming — a method where the computer itself finds the best map shape by iterating through options and stress-testing them with built-in certificates.
The most interesting part is that the same approach can be adapted to tackle other open questions in quantum information theory. For instance, it can test the PPT-square conjecture (which states that the square of any trace-preserving map remains positive under partial transpose) or refine recently proposed bounds for eigenvalues of 2-positive maps. In essence, the scientists have created not just a set of new maps, but a whole factory for producing them — detectors for various dimensions and structural constraints can roll off the assembly line.
Let's return to our dice. If before we could only guess whether they were cheating, now we have a device that looks at the dice from an angle that makes any secret synchronization as clear as day. And this device can be recalibrated for card games, roulette, or twenty-sided dice: there are no more limits.
as written by the authors arXiv:2605.14644
Positive maps that are not decomposable are a key resource in entanglement theory because they can detect bound entangled states, yet systematic methods for constructing them remain limited. We introduce an optimization framework based on differentiable semidefinite programming (SDP) for generating positive non-decomposable maps under flexible structural constraints on their Choi matrices. The method combines SDP-based certificates of non-decomposability and positivity with gradient-based optimization, enabling a systematic search over maps with different input and output dimensions. Within this framework, we generate previously unknown numerical examples, identify a parametrized family of maps arising from masked Choi matrices, and construct real non-decomposable maps. We further show that the same approach can be adapted to explore open questions in quantum information theory, including the PPT square conjecture and recently proposed eigenvalue bounds for 2-positive trace-preserving maps.
Standard methods only see entanglement if it's "loud"—like a shout in a quiet room. But non-decomposable maps can hear even a whisper: they can detect so-called bound entanglement, which cannot be used directly but can be "purified" into full-fledged entanglement.
In Hannu Rajaniemi's novel "The Quantum Thief," entanglement is used as currency and weapon, and its hidden forms serve as perfect hiding places, invisible to ordinary scanners. The new method is precisely such an advanced scanner.