Speakers
Description
Imagine handing a sealed box to someone, asking them to perform the quantum Fourier transform (QFT)—the heart of many quantum algorithms—and then returning the box to you without ever peeking inside. We show that this can be done with almost no extra cost or communication. The trick has two layers:
-
Quantum cloak. Before sending the state, the client scrambles it with a randomly chosen Pauli operator (a “quantum one-time pad”). To anyone who lacks the key, the qubits look completely thermal.
-
Classical bookkeeping. The two numbers that specify that Pauli mask are hidden in ordinary RSA ciphertexts—a public-key code. Although the server never learns the key, the very structure of the QFT causes the hidden mask to evolve linearly as the circuit runs. Linear evolution is precisely what RSA can update blindly: the server just multiplies the two ciphertexts by known constants after each gate while applying the usual QFT operations on the qubits.
Because the Fourier network is built entirely from Clifford gates, the procedure needs no exotic resources: no interaction, no “magic states,” and no heavyweight lattice cryptography. The server expends the same O((log N)^2) gate count it would on an unprotected register, while the client finishes with a single RSA inversion and one Pauli correction.
The result offers a fresh perspective on a long-standing puzzle in fundamental physics. If black holes process infalling matter through unitary dynamics that are effectively Clifford-only, their outgoing Hawking radiation would look maximally mixed to an external observer—exactly as we see—while still encoding the interior computation behind a quantum one-time pad. In that picture the event horizon acts as the untrusted server, matter fields supply the data, and spacetime itself provides the passive RSA-like bookkeeping that lets the hidden key propagate coherently. What appears to be random noise could therefore be the ciphertext of a vast homomorphic computation, preserving unitarity without revealing the underlying information.
Details
-
| Internet talk | No |
|---|---|
| Is this an abstract from experimental collaboration? | No |
| Name of experiment and experimental site | - |
| Is the speaker for that presentation defined? | No |