toplogo
Sign In

Unconditional Quantum Advantage for Sampling with Shallow Circuits


Core Concepts
Quantum circuits can sample from distributions that classical circuits cannot, even with shallow depths.
Abstract
Recent research shows that constant-depth quantum circuits can sample from distributions that are hard for classical circuits to reproduce. The study focuses on input-independent sampling tasks and provides unconditional proofs of quantum advantage over classical methods. By utilizing GHZ states and non-unitary gates, the research demonstrates the ability of quantum circuits to approximate complex distributions efficiently. The results highlight the potential of quantum computing in solving sampling problems beyond the capabilities of classical systems.
Stats
For each δ < 1, there exists a family of distributions {Dn} such that Cn samples from Dn within total variation distance at most 1/6 + O(n^-c). Any classical circuit with fan-in 2 taking n + nδ random bits has depth Ω(log log n) to produce output close to Dn. The distribution (X, majmodp(X)) is hard to sample from for low-depth classical circuits with bounded fan-in and limited inputs. Viola's techniques prove lower bounds for locality functions related to circuit depth. A Poor Man’s GHZ state allows for a constant-depth quantum circuit sampling distribution close to (X, majmodp(X) ⊕ parity(X)).
Quotes
"Quantum devices can process information in ways that classical devices cannot." "Constant-depth quantum circuits can sample from distributions that can't be reproduced by constant-depth bounded fan-in classical circuits." "The complexity of state preparation has implications in quantum cryptography and physics."

Key Insights Distilled From

by Adam Bene Wa... at arxiv.org 03-19-2024

https://arxiv.org/pdf/2301.00995.pdf
Unconditional Quantum Advantage for Sampling with Shallow Circuits

Deeper Inquiries

How many samples are needed to verify the distribution produced by a quantum circuit against all possible outputs

To verify the distribution produced by a quantum circuit against all possible outputs, only a few samples are needed due to the constant total variation distance mentioned in Corollary 4. This means that verifying the distribution produced by the quantum circuit requires minimal sampling. However, ruling out all distributions producible by classical circuits is a more challenging task and would require additional analysis.

Can the compilation technique for unitary gates be optimized for easier experimental implementation

The compilation technique for unitary gates can potentially be optimized for easier experimental implementation. By carefully considering alternative compilation methods, it may be possible to find a more natural approach that reduces the number of elementary gates required. Finding such an optimization could significantly enhance the feasibility of implementing the circuits described in this paper experimentally.

Is it possible to extend Viola's techniques to prove an input-independent sampling separation between QNC0 and AC0 circuits

Extending Viola's techniques to prove an input-independent sampling separation between QNC0 and AC0 circuits is an intriguing possibility. If successful, this extension would provide a novel method for lower-bounding the circuit complexity of quantum states. By adapting these techniques to QNC0 circuits, we could potentially establish a sampling separation that holds regardless of input variations, offering valuable insights into quantum computing capabilities compared to classical counterparts like AC0 circuits.
0