The Pursuit of Uniqueness: Extending Valiant-Vazirani Theorem to the Probabilistic and Quantum Settings

Valiant-Vazirani showed in 1985 [VV85] that solving NP with the promise that “yes” instances have only one witness is powerful enough to solve the entire NP class (under randomized reductions). We are interested in extending this result to the quantum setting. We prove extensions to the classes Merlin-Arthur MA and Quantum-Classical-Merlin-Arthur QCMA. Our results have implications for the complexity of approximating the ground state energy of a quantum local Hamiltonian with a unique ground state and an inverse polynomial spectral gap. We show that the estimation (to within polynomial accuracy) of the ground state energy of poly-gapped 1-D local Hamiltonians is QCMA-hard [AN02], under randomized reductions. This is in stark contrast to the case of constant gapped 1-D Hamiltonians, which is in NP [Has07]. Moreover, it shows that unless QCMA can be reduced to NP by randomized reductions, there is no classical description of the ground state of every poly-gapped local Hamiltonian that allows efficient calculation of expectation values. Finally, we discuss a few of the obstacles to the establishment of an analogous result to the class Quantum-Merlin-Arthur (QMA). In particular, we show that random projections fail to provide a polynomial gap between two witnesses.

💡 Research Summary

**
The paper investigates whether the celebrated Valiant‑Vazirani theorem—originally showing that NP can be reduced to its unique‑witness version (UP) via randomized reductions—can be extended to probabilistic and quantum complexity classes. The authors succeed in generalizing the theorem to Merlin‑Arthur (MA) and Quantum‑Classical‑Merlin‑Arthur (QCMA), establishing that the unique‑witness promise problems UMA and UQCMA are as hard as the full classes MA and QCMA respectively under randomized polynomial‑time reductions.

The technical core revisits the classic Valiant‑Vazirani construction in three refinements. First, a naïve approach guesses the exact size of the accepting witness set and filters each candidate with probability 1/w, where w is the guessed size. Second, the authors observe that an approximate multiplicative estimate of w suffices, reducing the number of guesses from exponential to linear in the witness length. Third, they replace truly random subsets with pairwise‑independent hash functions, which have polynomial‑size descriptions and can be evaluated efficiently, thereby ensuring the reduction runs in randomized polynomial time.

When adapting this scheme to MA and QCMA, a new obstacle appears: “yes” instances may contain exponentially many witnesses in the “gap” interval (acceptance probability between 1/3 and 2/3) in addition to those in the high‑probability interval (>2/3). Randomly selecting witnesses would almost always pick from the gap, failing to isolate a unique high‑probability witness. The authors overcome this by partitioning the gap interval into polynomially many sub‑intervals, estimating the number of witnesses in each sub‑interval, and applying the hash‑filter to each. By guessing approximate sizes for both the high‑probability and gap sub‑intervals, they achieve a constant‑probability success of filtering exactly one witness that lies in the high‑probability region. This yields Theorem 2 (UMA is MA‑hard) and Theorem 3 (UQCMA is QCMA‑hard).

The paper then explores physical implications for local Hamiltonian problems. A k‑local Hamiltonian H = Σ_j H_j acting on n qubits is given with a promised spectral gap Δ ≥ 1/poly(n). For constant‑gap 1‑D Hamiltonians, Hastings showed that ground states admit efficient Matrix‑Product‑State (MPS) approximations, placing the problem in NP. The authors define a “unique” 1‑D Hamiltonian problem where the ground state is unique and all other eigenvalues lie above a higher threshold. They prove this problem is UQMA‑complete (Theorem 5) and, via the earlier QCMA‑hardness result, conclude that estimating the ground‑state energy of poly‑gapped 1‑D Hamiltonians is QCMA‑hard under randomized reductions (Corollary 6).

A striking corollary follows: if every poly‑gapped 1‑D Hamiltonian’s ground state could be approximated to inverse‑polynomial accuracy by a description using only polynomially many parameters (e.g., an MPS) that also permits efficient expectation‑value computation, then QCMA would be contained in RP·NP. Since such a containment is considered highly unlikely, the result provides a strong “no‑go” statement: poly‑gapped 1‑D Hamiltonians likely lack efficient classical descriptions of their ground states.

Finally, the authors address the extension to Quantum‑Merlin‑Arthur (QMA). They construct a family of QMA “yes” instances where the verifier accepts an entire two‑dimensional subspace V with probability 1 and rejects its orthogonal complement. Attempting a quantum analogue of the classical hash‑filter—random projection onto a subspace—fails: with overwhelming probability, all states in V have acceptance probabilities that differ by at most O(1/√N), an exponentially small gap. Even more sophisticated tools such as random von‑Neumann measurements or quantum t‑designs only guarantee a constant total variation distance between outcome distributions, which does not translate into an efficiently distinguishable gap (this problem is SZK‑complete). Consequently, a randomized reduction from QMA to its unique‑witness version UQMA remains open, and new techniques are required.

In summary, the paper successfully extends the Valiant‑Vazirani paradigm to MA and QCMA, derives profound consequences for the complexity of poly‑gapped 1‑D Hamiltonian ground‑state estimation, and clarifies why the quantum case (QMA) resists a similar treatment. The work bridges complexity theory and quantum many‑body physics, highlighting both achievable reductions and fundamental barriers that shape our understanding of quantum verification and Hamiltonian complexity.