Qudit-ADAPT-VQE: an adaptive variational algorithm with counterdiabatic-inspired improvements for qudits
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.
Kai: I'm Kai.
Mira: And I'm Mira, here with Lev.
Kai: We're looking at the paper: "Qudit-ADAPT-VQE: an adaptive variational algorithm with counterdiabatic-inspired improvements for qudits".
Mira: Here's the gist of the paper. Variational quantum algorithms based on qudits have attracted significant attention in recent years In this work, we propose to address these issues through a qudit implementation of the ADAPT-VQE algorithm, which constructs the ansatz iteratively We show that the warm-start strategy inherent to ADAPT-VQE, together with an ansatz construction based on counterdiabatic operators, achieves higher accuracy and lower native gates implementation than approaches on fixed ansatz Furthermore, we show that, in qudit-based quantum computing, ADAPT-VQE with a counterdiabatic operator pool can navigate rough optimization landscapes with local traps through the burrowing mechanism, suggesting robustness against barren plateau effects and providing a scalable framework for variational quantum algorithms with qudits Qudits offer several potential advantages over the conventional qubit-based counterpart because each qudit is described by a d-dimensional local Hilbert space, with d > 2 This reduction may lower circuit complexity and improve resource efficiency, which is particularly relevant in the Noisy Intermediate Scale Quantum (NISQ) regime Moreover, qudits provide a natural representation for problems formulated in variables with dimensionality higher than binary, such as Max k-Cut Adaptive Ansatz Construction The ADAPT-VQE algorithm constructs the ansatz iteratively by selecting, at each step, the operator that provides the largest energy gradient from a predefined operator pool This procedure yields a problem-tailored ansatz that incorporates only the most relevant directions in Hilbert space, reducing circuit depth and avoiding unnecessary variational parameters The algorithm is initialized at iteration k = 0 with the state ϕgE because it is the ground state of HM At each iteration k, the energy gradient associated with each operator Vj is evaluated at θj = 0 as g(k)j=∂E/∂θj If g(k)2 < ε, the algorithm terminates, where ε is a predefined convergence threshold Counterdiabatic Improvement Mechanism The operator pool is obtained from an approximate counterdiabatic Hamiltonian associated with an adiabatic evolution ending at the cost Hamiltonian termed Counterdiabatic ADAPT-VQE (CD-ADAPT) This construction has been numerically shown for electronic structure problems that accelerates convergence toward the ground state while preserving the adaptive ansatzbuilding strategy of ADAPT-VQE The operator pool V is used as the operator pool for CD-ADAPT because its operators retain their connection to the approximate AGP, as discussed in Appendix A Performance Comparison and Robustness Qudit-ADAPT generally achieves higher approximation ratios than qudit QAOA and, in several cases, requires fewer variational parameters Qudit-ADAPT also requires considerably fewer variational parameters and native gates to achieve a given level of accuracy than qudit QAOA The warm-start strategy used in Qudit-ADAPT leads to more consistent convergence This behavior is consistent with the “burrowing” mechanism identified for ADAPT-VQE, where the addition of a new operator can deform the variational landscape and lower the local minimum reached by the previous ansatz The Qudit-ADAPT framework is regarded as a flexible framework for implementing variational quantum algorithms on qudit architectures, addressing two central challenges of fixed-ansatz approaches: ansatz design and robustness against unfavorable optimization landscapes, including local minima and barren plateaus The combination of warmstart initialization, iterative ansatz construction, and a sufficiently expressive operator pool provides robustness against local minima and may also mitigate optimization difficulties associated with barren plateaus Hardware Implementation Considerations The entire comparison between the two bases is contained in g The Qudit-ADAPT ansatz is less costly to implement at the circuit level A circuit of p layers carries 2p parameters, and an odd budget is matched by appending a trailing cost half-layer The QAOA cost layer is compiled in the Gell-Mann form of Eq.
Title and authors: Kai: The title itself is quite dense, but it tells you right away that they are taking ADAPT-VQE, which is already adaptive, and making it better for qudits by adding counterdiabatic ideas.
Mira: That counterdiabatic part is the core improvement they are proposing—it’s like giving the algorithm a smart map to follow when it’s searching through the optimization landscape.
Lev: The authors, Molina, Díaz-Moraga, Goyeneche, and Tancara, are tackling this problem using qudits to solve problems that have higher dimensionality than binary choices.
Kai: Since qudits can naturally represent variables with three or more states—like ternary variables in an optimization problem—they offer a way to encode the problem directly without having to map it awkwardly onto qubits first.
Mira: That dimensionality is what gives them the potential for better resource efficiency, which is what they are aiming for when they compare this against qubit-based methods.
Lev: The implication is that if you can build an algorithm that's inherently more efficient because of the state space it uses, it could be much more practical for NISQ devices than trying to force a complex problem onto qubits.
Kai: So, think of it as finding a better language for describing the problem itself, rather than just translating the problem into binary states.
Mira: Right. And this paper is showing that this new approach doesn't just work theoretically; they’ve actually implemented it to solve Max three-Cut and have seen results <ref:2608.14981#pg1,it to solve Max 3-Cut>.
The paper's summary: Kai: The summary boils down to this: they take ADAPT-VQE, which picks the best operator at each step based on the energy gradient, and they enhance that operator pool using something inspired by counterdiabatic driving during adiabatic evolution.
Mira: So, instead of just picking operators based on a simple local gradient, they’re selecting operators from a pool designed to guide them along a path that avoids getting stuck in bad local minima.
Lev: The key is that this iterative process, combined with the warm-start technique and this counterdiabatic operator construction, makes the algorithm more robust against those rough optimization landscapes we talked about before.
Kai: They claim that by adding one operator at a time and reusing previous results in a warm-start way, ADAPT-VQE stays in regions where the gradients are clear, allowing it to burrow through traps toward the actual solution.
Mira: It’s like having a better GPS for your search; instead of just blindly following the steepest downhill path, you have this informed pool of directions that steers you around roadblocks.
Lev: The paper applies this to solve Max three-Cut, and they show that Qudit-ADAPT generally beats qudit QAOA in terms of approximation ratios and requires fewer variational parameters overall <ref:2608.14981#pg1,to solve Max 3-Cut>.
Kai: So the practical implication is that if you’re designing a new algorithm for optimization problems on quantum hardware, this framework gives you a proven way to build something that’s both accurate and resource-efficient.
The paper's improvements: Mira: The main improvement they highlight is using that counterdiabatic operator pool, which is built from an approximate counterdiabatic Hamiltonian, to guide the ansatz construction process.
Kai: This pool isn't just any set of operators; it’s specifically constructed to accelerate convergence toward the ground state while keeping the adaptive ansatz strategy intact.
Lev: They use this pool because its operators still have a connection to an approximate adiabatic gauge potential, which helps ensure that the construction process is mathematically sound and related to known physical processes.
Kai: And they also mention extending this idea by considering higher-order approximations of that approximate adiabatic gauge potential, which lets them systematically increase the expressivity of their operator pool.
Mira: That increased expressivity is what lets them achieve approximation ratios above zero point nine nine for most instances they test on, which speaks to how powerful this guided construction can be <ref:2608.14981#pg1>.
Lev: From a hardware perspective, there’s also an improvement to consider—they suggest selecting specific groups of commuting operators from that pool.
Kai: That would help reduce the measurement cost associated with evaluating those gradients during the adaptive step, which is a real concern when you start thinking about running this on real hardware.
Mira: So they’re not just looking at accuracy; they’re also thinking about how to make the actual execution of this algorithm cheaper and more manageable for real quantum chips.
Conclusion: Kai: So, to wrap up on "Qudit-ADAPT-VQE: an adaptive variational algorithm with counterdiabatic-inspired improvements for qudits," this framework is flexible because it combines warm-start initialization with iterative construction guided by counterdiabatic operators.
Mira: It addresses the two big hurdles for fixed ansatz methods: designing a good ansatz and dealing with those tricky optimization landscapes like local minima or barren plateaus.
Lev: I just want to say that while the theoretical gains are strong, you have to remember that running this on real hardware means the overhead from those gradient evaluations has to be managed carefully, especially if you're aiming for high fidelity.
Kai: Yeah, and we need to see how well this scales when we move past Max three-Cut problems into more complex scenarios <ref:2608.14981#pg1>.
Mira: It’s a solid framework for variational quantum algorithms on qudit architectures, showing that adaptive construction plus counterdiabatic inspiration can lead to better performance than fixed ansatz approaches.
Lev: Yeah, it’s a scalable starting point because it shows the logic works across different problem types, even if we still need to refine the hardware implementation details for practical use.
Joaquín Molina, Herbert Díaz-Moraga, Dardo Goyeneche, Diego Tancara
Facultad de Física, Pontificia Universidad Católica de Chile
quant-ph
Submitted: 2026-08-15
Updated: 2026-11-12
Comments: 14 pages, 6 figures
Code: https://github.com/JoaquinQu/Qudit-ADAPT
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 83/100
The gist: Variational quantum algorithms based on qudits have attracted significant attention in recent years In this work…
Key concepts
- Qudits
- Qudits are quantum systems that use a Hilbert space with dimension $d > 2$, unlike standard qubits ($d=2$). This higher dimensionality can potentially reduce circuit complexity and improve resource efficiency for certain problems, especially in the NISQ era.
- ADAPT-VQE
- ADAPT-VQE is an iterative algorithm that constructs a quantum ansatz by selecting the operator from a predefined pool that yields the largest energy gradient at each step. This creates a problem-tailored ansatz, focusing only on relevant directions in Hilbert space to reduce circuit depth.
- Counterdiabatic Operators
- These operators are derived from an approximate counterdiabatic Hamiltonian associated with adiabatic evolution. They form the operator pool for CD-ADAPT, helping the algorithm navigate complex energy landscapes more effectively and accelerate convergence toward the ground state.
Terminology
Summary
Variational quantum algorithms based on qudits have attracted significant attention in recent years In this work, we propose to address these issues through a qudit implementation of the ADAPT-VQE algorithm, which constructs the ansatz iteratively We show that the warm-start strategy inherent to ADAPT-VQE, together with an ansatz construction based on counterdiabatic operators, achieves higher accuracy and lower native gates implementation than approaches on fixed ansatz Furthermore, we show that, in qudit-based quantum computing, ADAPT-VQE with a counterdiabatic operator pool can navigate rough optimization landscapes with local traps through the burrowing mechanism, suggesting robustness against barren plateau effects and providing a scalable framework for variational quantum algorithms with qudits
Introduction to Qudit Advantages
Qudits offer several potential advantages over the conventional qubit-based counterpart because each qudit is described by a d-dimensional local Hilbert space, with d > 2 This reduction may lower circuit complexity and improve resource efficiency, which is particularly relevant in the Noisy Intermediate Scale Quantum (NISQ) regime Moreover, qudits provide a natural representation for problems formulated in variables with dimensionality higher than binary, such as Max k-Cut
Adaptive Ansatz Construction
The ADAPT-VQE algorithm constructs the ansatz iteratively by selecting, at each step, the operator that provides the largest energy gradient from a predefined operator pool This procedure yields a problem-tailored ansatz that incorporates only the most relevant directions in Hilbert space, reducing circuit depth and avoiding unnecessary variational parameters The algorithm is initialized at iteration k = 0 with the state ϕgE because it is the ground state of HM At each iteration k, the energy gradient associated with each operator Vj is evaluated at θj = 0 as g(k)j=∂E/∂θj If g(k)2 < ε, the algorithm terminates, where ε is a predefined convergence threshold
Counterdiabatic Improvement Mechanism
The operator pool is obtained from an approximate counterdiabatic Hamiltonian associated with an adiabatic evolution ending at the cost Hamiltonian termed Counterdiabatic ADAPT-VQE (CD-ADAPT) This construction has been numerically shown for electronic structure problems that accelerates convergence toward the ground state while preserving the adaptive ansatzbuilding strategy of ADAPT-VQE The operator pool V is used as the operator pool for CD-ADAPT because its operators retain their connection to the approximate AGP, as discussed in Appendix A
Performance Comparison and Robustness
Qudit-ADAPT generally achieves higher approximation ratios than qudit QAOA and, in several cases, requires fewer variational parameters Qudit-ADAPT also requires considerably fewer variational parameters and native gates to achieve a given level of accuracy than qudit QAOA The warm-start strategy used in Qudit-ADAPT leads to more consistent convergence This behavior is consistent with the “burrowing” mechanism identified for ADAPT-VQE, where the addition of a new operator can deform the variational landscape and lower the local minimum reached by the previous ansatz
Conclusion on Framework Utility
The Qudit-ADAPT framework is regarded as a flexible framework for implementing variational quantum algorithms on qudit architectures, addressing two central challenges of fixed-ansatz approaches: ansatz design and robustness against unfavorable optimization landscapes, including local minima and barren plateaus The combination of warmstart initialization, iterative ansatz construction, and a sufficiently expressive operator pool provides robustness against local minima and may also mitigate optimization difficulties associated with barren plateaus
Hardware Implementation Considerations
The entire comparison between the two bases is contained in g The Qudit-ADAPT ansatz is less costly to implement at the circuit level A circuit of p layers carries 2p parameters, and an odd budget is matched by appending a trailing cost half-layer The QAOA cost layer is compiled in the Gell-Mann form of Eq.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Adaptive ansatz construction for qudits can navigate rough optimization landscapes by selecting operators based on a gradient criterion from an operator pool derived from approximate counterdiabatic driving, which suggests
robustness against barren plateau effects and providing a scalable framework for variational quantum algorithms with qudits.
-
Qudit-ADAPT can achieve higher accuracy with lower variational and circuit-resource requirements than qudit QAOA because the
warm-start strategy inherent to ADAPT-VQE, together with an ansatz construction based on counterdiabatic operators, achieves higher accuracy and lower native gates implementation than approaches on fixed ansatz.
-
The algorithm can be made hardware-aware by incorporating
hardware-aware penalties into the gradient-based operator-selection criterion
to reduce measurement cost associated with pool gradients. -
By considering higher-order approximations to the approximate adiabatic gauge potential, Qudit-ADAPT can systematically increase the expressivity of the operator pool, allowing it to achieve
approximation ratios above 0.99 for the majority of the instances.
-
The system can be optimized for specific hardware by selecting groups of commuting operators to reduce measurement cost associated with evaluating pool gradients, as suggested by
selecting groups of commuting operators to reduce the measurement cost associated with evaluating pool gradients.
Abstract
Variational quantum algorithms based on qudits have attracted significant attention in recent years due to their potential for more efficient quantum information encoding and processing. However, as in their qubit-based counterparts, challenges such as barren plateaus and the design of efficient ansatzes remain major obstacles. In this work, we propose to address these issues through a qudit implementation of the ADAPT-VQE algorithm, which constructs the ansatz iteratively. Specifically, we introduce an operator pool inspired by adiabatic evolution enhanced with counterdiabatic driving for ansatz construction and employ it to solve the Max 3-Cut and Multiway Number Partitioning problems. We show that the warm-start strategy inherent to ADAPT-VQE, together with an ansatz construction based on counterdiabatic operators, achieves higher accuracy than fixed-ansatz approaches and, at equal number of variational parameters, requires fewer native gates. Furthermore, we show that, in qudit-based quantum computing, ADAPT-VQE with a counterdiabatic operator pool can navigate rough optimization landscapes with local traps through the burrowing mechanism, suggesting robustness against local minima and providing a flexible framework for variational quantum algorithms with qudits.
Sources
- A Quantum Approximate Optimization Algorithm
- Co-Designed Adaptive Quantum State Preparation Protocols
- How to really measure operator gradients in ADAPT-VQE
Related papers
- Computational regimes in matrix-product-state-based quantum trajectory simulations
- Toward chemically accurate quantum simulations on quantum hardware: A quantum-HPC hybrid approach
- Transversal surface-code game with reconfigurable qubits
- Fulqrum: a generalized framework for quantum subspace diagonalization
- Flux-Activated Resonant Control of a Bosonic Quantum Memory
- Machine learning the effects of many quantum measurements