This paper, recently published by researchers at Los Alamos National Laboratory and UT Austin, tackles one of the most frustrating bottlenecks in experimental quantum physics: the post-selection barrier.
TL;DR
Researchers have developed a deterministic method to prepare specific "quantum trajectories"—sequences of measurement outcomes—in polynomial time, bypassing the exponential cost usually required to observe rare quantum states. By leveraging imaginary time evolution, the team provides a roadmap for studying measurement-induced phase transitions at scales previously thought to be experimentally unreachable.
What Problem Does This Paper Address?
In quantum mechanics, measurements are inherently random. If you measure a qubit, you might get a 0 or a 1. If you have a circuit with 50 mid-circuit measurements and you want to study one specific sequence of outcomes (a "trajectory"), the odds of that specific sequence occurring naturally are $1$ in $2^{50}$.
To gather enough data to understand the physics of that single trajectory, you would have to run the experiment an astronomical number of times. This is the post-selection barrier. It recently forced researchers using IBM’s quantum computers to spend seven months running a 14-qubit experiment—a task that should have been routine. This paper asks: Can we "force" the quantum computer to give us the trajectory we want without waiting for luck?
Methods
The authors propose using Deterministic Quantum Imaginary Time Evolution (DQITE).
- Imaginary Time Evolution (ITE): Mathematically, applying an operator $e^{-\beta H}$ to a quantum state projects it onto the ground state of the Hamiltonian $H$. If you design $H$ such that its ground state is your desired measurement outcome (e.g., $|0\rangle$), ITE will "push" the state toward that outcome.
- The Unitary Trick: ITE is "non-unitary," meaning it doesn't preserve the total probability and can't be directly run on a quantum computer. However, the authors use the DQITE algorithm, which finds a specific unitary (a valid quantum gate) that mimics ITE for a specific input state.
- Locality: The method relies on the fact that in certain phases of matter, quantum correlations are "short-range." This allows the algorithm to work locally on small clusters of qubits rather than the whole system at once, keeping the computational cost low (polynomial scaling).
Datasets
As a theoretical physics paper, the "datasets" consist of numerical simulations of many-body quantum systems:
- Random Quantum Circuits: Brickwork layers of two-qubit Clifford and Haar-random gates.
- System Sizes: Simulations were performed on systems ranging from 20 qubits (for full DQITE validation) up to 256 qubits (to analyze correlation decay).
- Benchmarks: The results were compared against the known critical threshold for measurement-induced phase transitions ($p_c \approx 0.16$).
Key Insights and Ideas
- The Area Law Advantage: The algorithm's efficiency is tied to the entanglement structure of the state. It works perfectly in the "area law" phase (where measurements are frequent), because correlations decay exponentially. In the "volume law" phase (low measurement rate), the long-range entanglement breaks the local assumptions of the algorithm.
- Complexity Limits: The authors prove that a "universal" version of this algorithm (one that works for any state) is likely impossible. If it existed, it would imply that quantum computers could solve PP-complete problems (a massive class of hard problems including NP), which would collapse modern complexity theory ($BQP = PP$).
- Physical Observation vs. Manual Choice: A key insight is that we don't need to prepare any arbitrary trajectory. We only need to prepare trajectories that could have happened naturally. This "physically motivated" constraint is what makes the polynomial-time solution possible.
Key Results
- Exponential Error Suppression: The fidelity of the prepared state to the target state improves exponentially as the "imaginary time" ($\beta$) increases.
- Polynomial Scaling: The total runtime for preparing a trajectory with $M$ measurements scales as $O(M^2)$, a massive improvement over the $2^M$ brute-force approach.
- Phase Separation: Numerical results (Fig. 4) show a clear distinction: in the area-law regime, the algorithm converges rapidly to high fidelity, while in the volume-law regime, it fails as expected, confirming it as a tool specifically for the area-law phase.
Limitations and Future Work
- The Volume Law Barrier: The method currently cannot handle states with high, long-range entanglement (the volume law phase). Finding a way to probe the transition point itself remains a challenge.
- Hardware Noise: While the algorithm is "efficient" in terms of gate count, it still requires high-fidelity operations. On current "Noisy Intermediate-Scale Quantum" (NISQ) devices, the accumulation of errors during the DQITE steps might still be a hurdle.
- Beyond MIPT: The authors suggest this deterministic post-selection could be applied to other fields, such as "deep thermalization," where preparing specific rare states is currently the primary experimental bottleneck.
Summary updated