Pith. sign in

REVIEW 5 major objections 8 minor 45 references

Quantum Agents

T0 review · 5 major / 8 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read The paper defines a quantum agent as an autonomous system whose reasoning, decision-making, learning, or sensing is enhanced by quantum operations, formalised as the tuple (Q,C,M,P,A), and demonstrates the definition with three simulated…

desk verdict A useful vocabulary and maturity model for a young subfield, but the prototype evidence does not substantiate the feasibility claim. read the letter →

arxiv 2506.01536 v2 pith:F34UILCN submitted 2025-06-02 quant-ph

classification quant-ph MSC 81P6868T4268Q1268T05
keywords quantumagentsagenticAIreinforcementlearningvariationalcircuitsGroversearchimageencryptionagentmaturitymodelNISQ
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper attempts to give the emerging field of quantum-agentic AI a shared, formal foundation. It defines a quantum agent as an autonomous system whose reasoning, decision-making, learning, or sensing is enhanced by quantum operations, captured by the tuple (Q,C,M,P,A): quantum processing resources, classical control logic, hybrid memory, perception, and action. It then places agents on a four-level maturity ladder, from NISQ-optimised decision agents to fully quantum-native systems. Three prototype agents — a Grover-search action selector, a variational-circuit multi-armed bandit, and an adaptive image-encryption agent — are offered as evidence that such systems can already be built and run in simulation. A sympathetic reader would take the paper's contribution to be the definition and vocabulary, with the prototypes as existence proofs.

What carries the argument

The load-bearing object is the tuple (Q,C,M,P,A), read as the anatomy of a quantum agent: quantum processing resources, classical control logic, hybrid memory, perception, and action. It carries the argument by giving every later component — maturity levels, architectures, and prototypes — a common vocabulary; each maturity level is specified by how each tuple element evolves, from NISQ-scale parameterized circuits at Level 1 to a fault-tolerant universal QPU at Level 4. The prototypes instantiate the tuple in two-qubit circuits: a Grover oracle-plus-diffuser for action selection, a variational policy circuit trained by gradient descent for the bandit, and a variational policy selecting among XOR, QFT, and scrambling for encryption.

What would settle it

Run the bandit agent or the Grover agent on a real noisy quantum device with the same circuits and compare success rates and cumulative reward against the simulator outputs and against a classical epsilon-greedy baseline; if noisy execution falls to random-choice level or below the classical baseline, the paper's demonstration of NISQ-era feasibility fails.

Watch

Extended reading notes

Core claim

The paper's central claim is its formal definition: a quantum agent is an autonomous system characterized by the tuple (Q,C,M,P,A), where Q is the set of quantum processing resources, C is classical control logic, M is hybrid memory, P is perception, and A is action, and where the defining feature is that quantum operations enhance or enable the agent's reasoning, decision-making, learning, or sensing. The paper also claims that this definition supports a four-level maturity model aligned with the expected evolution of quantum hardware, and that the three prototypes demonstrate the feasibility of this framework in the NISQ era.

Load-bearing premise

The feasibility claim rests on the assumption that behaviour observed in a noiseless two-to-four qubit simulator transfers to real NISQ hardware, and that the chosen reward signals (cumulative reward, ciphertext entropy) faithfully measure real-world value.

Editorial extensions

If this is right

  • If the definition holds, researchers and developers get a common language for describing and comparing quantum-agentic systems, which is a precondition for benchmarking.
  • If the prototypes transfer to hardware, then search-based action selection, quantum reinforcement learning, and adaptive quantum encryption become practical targets for NISQ-era deployment.
  • The maturity model implies a staged roadmap: agents can be designed for today's devices at Level 1 while architectures can be prepared for fault-tolerant hardware at Level 4.
  • The tuple-based account makes the bidirectional claim concrete: quantum operations can enhance agency, and agentic control can orchestrate quantum workflows, within one formal description.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The definition is broad enough that nearly any hybrid quantum-classical system with a feedback loop could qualify as a quantum agent; the sharper question is what counts as a non-classical advantage, which the paper leaves open.
  • A natural test is to run the same three prototypes on real noisy hardware with noise-model characterisation and against classical baselines such as epsilon-greedy bandits, comparing sample complexity and success rates.
  • The maturity model could be operationalised as a scoring rubric with verifiable hardware milestones — error rates, qubit counts, memory coherence — making a claimed maturity level a checkable statement rather than a label.
  • The encryption agent's entropy reward is a proxy; an adversarial test that attempts statistical attacks on the ciphertexts would show whether adaptive strategy selection actually improves security.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 8 minor

Summary. The paper proposes a formal definition of quantum agents as a tuple (Q,C,M,P,A), where Q is quantum processing resources, C is classical control logic, M is hybrid memory, P is perception, and A is action, with the defining condition that quantum operations provide non-classical advantages in the agent's reasoning, learning, or sensing. It further introduces a four-level maturity model (Levels 1–4) aligned with a DOE quantum development timeline, discusses quantum-assisted versus quantum-centric control modes, sketches architectures and use cases, and presents three prototype agents implemented with Qiskit and PennyLane: a two-qubit Grover-based action selector (Agent 1), a variational-circuit multi-armed bandit (Agent 2), and a variational-policy adaptive image encryption agent (Agent 3). The paper's central claim is that these prototypes demonstrate the feasibility of quantum-agentic systems in the NISQ era.

Significance. If the proposed framework were adopted, the paper would provide a shared vocabulary and a maturity model for an emerging interdisciplinary field, which is a useful conceptual contribution. The paper also correctly identifies open problems such as the need for benchmarks that isolate the contribution of quantum components. However, the experimental evidence for the stated feasibility claim is not sufficient: the prototypes are extremely simple, are run only on a noiseless statevector simulator, include no classical baselines, no error bars, no hardware execution, and no noise model. As a result, the paper's significance currently rests on its conceptual scaffolding rather than on any demonstrated quantum-agentic capability.

major comments (5)
  1. [Section 5.2 and Section 10] The paper's own defining criterion for a quantum agent is that reasoning, decision-making, learning, or sensing be enhanced by quantum operations that provide non-classical computational or informational advantages. None of the three prototypes in Section 10 is tested against this criterion. No classical control arms (e.g., a classical epsilon-greedy bandit, a fixed QFT encryption policy, or a classical oracle-based search) are reported, so the reader cannot determine whether the quantum circuits contribute anything beyond trivial classical behavior. Without such baselines, the claim in the abstract and Section 10 that the prototypes 'demonstrate the feasibility' of the framework is unsupported.
  2. [Section 10.1, Agent 1] Agent 1 is a fixed two-qubit Grover circuit with the oracle manually constructed to mark the target state |10>. At n=2, this is a textbook textbook circuit with no agentic loop, no decision feedback, and no measurable speedup; the number of oracle queries is fixed at one. The observed correctness of the output ('Quantum Agent chose action: 10') is trivially expected from the circuit's construction and does not illustrate 'search-based decision-making' as an agentic capability. This prototype therefore provides no evidence relevant to the feasibility of quantum agents.
  3. [Section 10.3, Agent 3] The reward function in Algorithm 3 is defined as the entropy of the encrypted image (line 22), and the policy is trained to maximize this entropy. Since QFT is known to increase ciphertext entropy, the observed QFT-dominant action histogram in Figure 16 is a direct consequence of the reward definition, not evidence of learned adaptivity. A classical controller that always selects QFT would achieve the same reward. The claim in Section 10.3.6 that the agent 'learns to prefer transformations that yield higher ciphertext entropy' is therefore circular: the training objective already encodes that preference. This prototype does not demonstrate quantum enhancement of the agent's decision-making.
  4. [Section 10.2, Agent 2] Agent 2 solves a four-armed bandit with reward probabilities 0.2–0.8 using a four-parameter variational quantum circuit as the policy. This is a trivial reinforcement-learning problem that a classical softmax or epsilon-greedy policy would solve with comparable or better performance. The paper reports no classical baseline, no statistical repetition over seeds, and no error bars; Figure 7 shows a cumulative reward curve that is not compared against any reference. Consequently, the conclusion that 'a variational quantum circuit can successfully serve as a learnable policy in reinforcement learning' does not establish anything specific to quantum agents, nor does it support the broader feasibility claim.
  5. [Section 11.2] Section 11.2 lists 'How do we isolate the contribution of quantum components from overall system behavior?' as an open question, yet Section 10 presents the prototypes as evidence of feasibility without applying any such isolation. This is an internal inconsistency: the paper acknowledges the evaluation challenge but then makes claims that presuppose its resolution. The authors should either add the missing controls and comparisons or explicitly reframe Section 10 as a demonstration of the architectural concepts only, with no claim of quantum advantage.
minor comments (8)
  1. [Abstract and Section 10] The abstract states that the prototypes 'demonstrate the feasibility' of the proposed framework, but the simulations are noiseless and hardware-free. We recommend softening this to 'illustrate the architectural concepts' or adding the necessary hardware and noise studies to support the original wording.
  2. [Section 7, Table 1] There is a terminology inconsistency: the text refers to a 'first level, ad-hoc Quantum Agent', but Table 1 labels Level 1 as 'NISQ-Optimized Decision Agents' and the maturity model in Figure 2 uses 'NISQ-Optimised Quantum Resilient Agent'. Please unify the nomenclature across the text, table, and figure.
  3. [Section 3.3] The sentence 'instead - Probabilistic nature of quantum calculation proves it potentially useful at simulating complex Stochastic processes.' appears to be an unfinished or misplaced editorial note. It should be either integrated into the preceding discussion or removed.
  4. [Section 4.2] There is a typo: 'following the advise advice of Richard Feynman' should read 'following the advice of Richard Feynman'.
  5. [Section 2.2] Reference [2] is described as demonstrating 'systematic quantum advantage' with 'faster learning compared to classical counterparts'. This is a strong claim about a Nature paper; please verify that the finding is correctly characterized and specify whether the advantage is provable or empirical.
  6. [References] Reference [6] is a placeholder 'A. Author and B. Author'. This is not acceptable in a submitted manuscript and must be replaced with a proper citation.
  7. [Section 10.2.1] The informal phrase 'Let's consider' is used in a formal paper; please rephrase (e.g., 'We consider').
  8. [Figures 7–9 and 15–16] Several figures (e.g., Figure 9 and Figure 16) are histograms without axis labels or legends. Please add full axis labels and, where appropriate, error bars or multiple-seed distributions to support the claimed learning behavior.

Circularity Check

1 steps flagged · score 2.0 of 10

One prototype's 'correct action' is hard-coded into the Grover oracle, making its success self-definitional; the central definition and other prototypes are otherwise independent.

  1. self definitional [Section 10.1.1 (Agent Design and Action Encoding) and Algorithm 1]
    "The agent is tasked with identifying the correct action, predefined as |10>... Algorithm 1: 1: Define the correct action as bitstring a* =10 ... 4: Construct oracle O such that O|a*>=−|a*> ... The Grover-based quantum agent correctly identified the optimal action (10) matching the environment's target, demonstrating reliable decision-making via quantum search."

    The target action is not inferred from the environment; it is set as an input (a* = 10) and inserted directly into the oracle. Grover's algorithm then amplifies exactly that pre-marked state, so the measured outcome '10' is guaranteed by construction. Reporting this as 'reliable decision-making' restates the input oracle as the output decision: the success criterion is identical to the construction parameter. This is a textbook oracle circuit rather than an agentic or predictive result, so it cannot independently demonstrate decision-making. The circularity is minor and confined to Agent 1; the formal definition and the other two prototypes are not reduced this way.

full rationale

The paper is primarily definitional and exploratory; there is no theorem whose conclusion is assumed. The formal tuple (Q,C,M,P,A) and the four-level maturity model are stipulated, not derived, and the DOE timeline is an external reference. The only concrete reduction-by-construction is Agent 1: the correct action is hard-coded into the Grover oracle, so its reported success is equivalent to its input. Agent 2 learns against a stochastic bandit, and Agent 3 optimizes a defined entropy reward; their results are weak evidence without classical baselines or hardware execution, but that is an experimental-support gap rather than circularity. Reference [25] is a self-citation by two co-authors but is used only as an illustrative encryption example and is not load-bearing. Score 2 reflects the one self-definitional prototype while recognizing that the central conceptual content is independent.

Assumptions & free parameters 2 free parameters · 4 assumptions · 3 invented entities

The paper's central definition is a conceptual artifact: the tuple (Q,C,M,P,A) and the four maturity levels are not derived from data or machine-checked. The only quantitative content lives in the three prototype agents, which fit variational weights to toy reward functions. The framework rests on standard quantum-mechanics postulates plus the field-level assumption that quantum speedups survive in end-to-end agentic loops.

free parameters (2)
  • Agent 2 variational weights theta_0..theta_3 = initialized to pi/4; Adam-updated over 100 episodes
    Trained to maximize cumulative reward in the toy four-armed bandit; the reported learning curve depends on these fitted weights.
  • Agent 3 variational weights theta_0..theta_3 = trained over 30 episodes to maximize ciphertext entropy
    The encryption-action policy is fit to the entropy reward; the reported preference for QFT follows from this objective rather than from independent evaluation.
assumptions (4)
  • standard math Standard postulates of quantum mechanics: unitary evolution, Born rule measurement probabilities, no-cloning, entanglement.
    Used implicitly in all prototype circuits (Sections 10.1-10.3) and in the memory discussion of Section 5.2.
  • domain assumption Quantum algorithms such as Grover, QAOA, QSVT, and quantum transformers retain their theoretical advantages when embedded as subroutines in an agent's perception-decision-action loop.
    Sections 3.1-3.3 and Section 7 assume these speedups transfer to end-to-end agentic workflows, a premise not demonstrated by the paper.
  • domain assumption The DOE quantum development timeline is accurate enough to calibrate the maturity model.
    Section 7 anchors maturity levels to the US DOE roadmap; if hardware milestones slip, the Level 3 and Level 4 descriptions lose their grounding.
  • ad hoc to paper Ciphertext entropy maximization is a sufficient proxy for encryption security.
    Agent 3's reward function uses entropy only; no correlation analysis, uniformity tests, or attack-resistance metrics are used to validate security (Sections 10.3.5-10.3.6).
invented entities (3)
  • Quantum agent formal tuple (Q,C,M,P,A)
    purpose: Serve as the paper's central definition of a quantum agent and as the basis for the maturity model.
    No falsifiable prediction or unique theorem follows from the tuple; it is a proposed formalism with no independent evidence outside the paper.
  • Four-level Quantum Agent maturity model (Levels 1-4)
    purpose: Classify agents by hardware era and capability.
    A taxonomy anchored to the DOE timeline; it is not derived from measurements in the paper and cannot be falsified as presented.
  • Quantum-centric control / quantum mesh as decision substrate
    purpose: Describe architectures where agency partially resides in an entangled quantum state.
    Speculative concept in Section 6.2; no mechanism or experiment is provided, so there is no independent evidence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Quantum Agents." pith.science (2026). https://pith.science/paper/F34UILCN

@misc{pith2026250601536,
  author       = {Pith},
  title        = {Pith review of: Quantum Agents},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F34UILCN}},
  note         = {Machine review of arXiv:2506.01536}
}
read the original abstract

This paper explores the intersection of quantum computing and agentic AI by examining how quantum technologies can enhance the capabilities of autonomous agents, and, conversely, how agentic AI can support the advancement of quantum systems. We analyze both directions of this synergy and present conceptual and technical foundations for future quantum-agentic platforms. Our work introduces a formal definition of quantum agents and outlines potential architectures that integrate quantum computing with agent-based systems. As a proof-of-concept, we develop and evaluate three quantum agent prototypes that demonstrate the feasibility of our proposed framework. Furthermore, we discuss use cases from both perspectives, including quantum-enhanced decision-making, quantum planning and optimization, and AI-driven orchestration of quantum workflows. By bridging these fields, we aim to chart a path toward scalable, intelligent, and adaptive quantum-agentic ecosystems.

Figures

Figures reproduced from arXiv: 2506.01536 by the authors.

Figure 1
Figure 1. The anatomy of a Quantum Agent System Ar￾chitecture: a modular framework combining classical logic, quantum operations, safety mechanisms, and external inter￾faces for intelligent, auditable agent behavior. Classical tasks use the Model Context Protocol (MCP) for compatibility with existing AI systems. Quantum tasks run on the Quantum Process￾ing Unit (QPU), which handles complex calculations. A Knowledge Base provi… view at source ↗
Figure 2
Figure 2. The anatomy of a Quantum Agent based on its maturity model. capabilities, example being Perception L1 vs. Perception L4. The subse￾quent sections will outline the specifics of the evolutionary process from the NISQ level to the fully quantum-native agent, and how the architectural components specified in figure 1 will evolve [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. Quantum Agent-L2 A Quantum Agent-Level 2 uses Hybrid Quantum Machine Learning to evaluate the explainability of the behavior of agentic AI and its adherence to Responsible AI [26].Using a Markov Chain model, the meta-cognition layer continuously provides a real-time set of states, actions, and the maximum Q-value as input to the meta-meta-cognition layer. This layer uses QML to generate various future scenarios fast… view at source ↗
Figures from the paper (13 more)
Figure 4
Figure 4. Figure 4: Grover Circuit for Quantum Agent Quantum Agent chose action: 10 Environment's correct action was: 10 Was the agent correct? [PITH_FULL_IMAGE:figures/full_fig_p028_4.png]
Figure 5
Figure 5. Figure 5: Output Decision of Grover Circuit Quantum Agent 10.2. Quantum Multi-Armed Bandit Agent Using Variational Pol￾icy Circuits. Here we build a quantum learning agent that applies quan￾tum variational circuits to solve the classical multi-armed bandit (MAB) problem. The obj…
Figure 6
Figure 6. Figure 6: Quantum policy circuit used by the agent [PITH_FULL_IMAGE:figures/full_fig_p030_6.png]
Figure 7
Figure 7. Figure 7: Cumulative reward across training episodes, indi￾cating learning behaviour [PITH_FULL_IMAGE:figures/full_fig_p030_7.png]
Figure 8
Figure 8. Figure 8: Per-episode reward. Spikes correspond to success￾ful rewards; fluctuations reflect stochastic environment. confirms that the agent successfully identified the optimal arm 10 as the most frequently chosen. This experiment demonstrates that a variational quantum circuit …
Figure 9
Figure 9. Figure 9: Histogram of arm selection frequency. The agent converges to arm 10, the optimal choice. efficiently learns the best action in a noisy reward landscape, reinforcing the viability of quantum-enhanced learning frameworks for adaptive decision￾making tasks. 10.3. Quantum …
Figure 10
Figure 10. Figure 10: Architecture of the Quantum Agent for Adaptive Quantum Image Encryption. 10.3.3. Agent Design and Action Encoding. The Quantum Policy Agent is implemented as a 2-qubit variational quantum circuit (VQC), serving as the decision-making core of the encryption pipeline. T…
Figure 11
Figure 11. Figure 11: Quantum Policy Agent circuit. It encodes en￾tropy into qubit rotations and produces a distribution over four encryption actions. samples one operation per image segment. The pseudocode algorithm of this QIE agent is given in Algorithm 3. The circuit is trained using r…
Figure 12
Figure 12. Figure 12: Quantum XOR encryption circuit with ancilla qubits. characteristics, particularly for high-entropy inputs where additional com￾plexity yields diminishing returns in XOR-based masking [PITH_FULL_IMAGE:figures/full_fig_p036_12.png]
Figure 13
Figure 13. Figure 13: QFT encryption circuit for image segments. Scrambling/Gate Permutation: Scrambling is achieved through a com￾bination of SWAP and Pauli-X gates to reorder and invert specific qubit states as shown in [PITH_FULL_IMAGE:figures/full_fig_p036_13.png]
Figure 14
Figure 14. Figure 14: Scrambling circuit applying swaps and targeted X gates. 10.3.5. Reinforcement Learning Integration. The agent is trained using a re￾inforcement learning loop with entropy maximization as the reward proxy. Over 30 episodes, the agent receives image entropy as input and…
Figure 15
Figure 15. Figure 15: Encrypted image entropy over episodes. The agent learns to favour transformations that yield high cipher￾text entropy [PITH_FULL_IMAGE:figures/full_fig_p038_15.png]
Figure 16
Figure 16. Figure 16: Frequencies of encryption operations selected during training. QFT dominates due to its entropy ampli￾fication. it learns from ciphertext outcomes and adjusts operation preferences ac￾cordingly. This framework is especially promising for use cases in quantum￾secured s…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

45 extracted references · 37 canonical work pages

  1. [6]

    Author and B

    A. Author and B. Author. Quantum artificial intelligence: A brief survey.K¨ unstliche Intelligenz, 2024

  2. [7]

    Enrique Solano. Agentic quantum computing at kipu quantum.LinkedIn post, 2024.https://www.linkedin.com/posts/enrique-solano-812631154_ agentic-quantum-computing-at-kipu-quantum-activity-7325134022114725889-PNbJ

  3. [1]

    Toward quantum computational agents.DFKI, 2003

    Matthias Klusch. Toward quantum computational agents.DFKI, 2003

  4. [2]

    Experimental quantum speed-up in reinforcement learning agents.Nature, 591(7849):229–233, 2021

    Valeria Saggio, Bastian E Asenbeck, Andreas Hamann, et al. Experimental quantum speed-up in reinforcement learning agents.Nature, 591(7849):229–233, 2021. 42 SULTANOW, TEHRANI, DUTTA, BUCHANAN & KHAN

  5. [3]

    Quantum adaptive agents with efficient long-term memories.Physical Review X, 12(1):011007, 2022

    Thomas J Elliott, Mile Gu, Andrew JP Garner, et al. Quantum adaptive agents with efficient long-term memories.Physical Review X, 12(1):011007, 2022

  6. [4]

    Energetic advan- tages for quantum agents in online execution of complex strategies.arXiv preprint arXiv:2503.19896, 2025

    Jayne Thompson, Paul M Riechers, Andrew JP Garner, et al. Energetic advan- tages for quantum agents in online execution of complex strategies.arXiv preprint arXiv:2503.19896, 2025

  7. [5]

    Quantum Multi-Agent Reinforcement Learning via Variational Quantum Circuit Design

    Won Joon Yun, Yunseok Kwak, Jae Pyoung Kim, et al. Quantum multi-agent reinforcement learning via variational quantum circuit design.arXiv preprint arXiv:2203.10443, 2022

  8. [8]

    Quantum mechanics helps in searching for a needle in a haystack

    Lov K Grover. Quantum mechanics helps in searching for a needle in a haystack. Physical review letters, 79(2):325, 1997

Show all 45 references
  1. [9]

    Grover adaptive search for constrained polynomial binary optimization.Quantum, 5:428, 2021

    Austin Gilliam, Stefan Woerner, and Constantin Gonciulea. Grover adaptive search for constrained polynomial binary optimization.Quantum, 5:428, 2021

  2. [10]

    Amplitude amplification for optimization via subdivided phase oracle

    Naphan Benchasattabuse, Takahiko Satoh, Michal Hajduˇ sek, and Rodney Van Me- ter. Amplitude amplification for optimization via subdivided phase oracle. In2022 ieee international conference on quantum computing and engineering (qce), pages 22–30. IEEE, 2022

  3. [11]

    Gener- alized grover search algorithm for arbitrary initial amplitude distribution.Lecture notes in computer science, pages 140–147, 1999

    David Biron, Ofer Biham, Eli Biham, Markus Grassl, and Daniel A Lidar. Gener- alized grover search algorithm for arbitrary initial amplitude distribution.Lecture notes in computer science, pages 140–147, 1999

  4. [12]

    A quantum approximate op- timization algorithm, 2014

    Edward Farhi, Jeffrey Goldstone, and Sam Gutmann. A quantum approximate op- timization algorithm, 2014

  5. [13]

    A review on quantum approximate optimization algorithm and its variants.Physics Reports, 1068:1–66, 2024

    Kostas Blekos, Dean Brand, Andrea Ceschini, Chiao-Hui Chou, Rui-Hao Li, Komal Pandya, and Alessandro Summer. A review on quantum approximate optimization algorithm and its variants.Physics Reports, 1068:1–66, 2024

  6. [14]

    Springer Science & Business Media, 2005

    Arnab Das and Bikas K Chakrabarti.Quantum annealing and related optimization methods, volume 679. Springer Science & Business Media, 2005

  7. [15]

    Hybrid quantum-classical search algorithms.ACM Transactions on Quantum Computing, 5(2):1–18, 2024

    Ansis Rosmanis. Hybrid quantum-classical search algorithms.ACM Transactions on Quantum Computing, 5(2):1–18, 2024

  8. [16]

    Quantum circuit ansatz: Patterns of abstraction and reuse of quantum algorithm design, 2024

    Xiaoyu Guo, Takahiro Muta, and Jianjun Zhao. Quantum circuit ansatz: Patterns of abstraction and reuse of quantum algorithm design, 2024

  9. [17]

    Review of ansatz designing techniques for variational quantum algo- rithms, 2022

    Junhan Qin. Review of ansatz designing techniques for variational quantum algo- rithms, 2022

  10. [18]

    Exploring entanglement and optimization within the hamiltonian variational ansatz.PRX Quantum, 1:020319, Dec 2020

    Roeland Wiersema, Cunlu Zhou, Yvette de Sereville, Juan Felipe Carrasquilla, Yong Baek Kim, and Henry Yuen. Exploring entanglement and optimization within the hamiltonian variational ansatz.PRX Quantum, 1:020319, Dec 2020

  11. [19]

    Towards near-term quantum simulation of materials.Nature Communications, 15(1):211, 2024

    Laura Clinton, Toby Cubitt, Brian Flynn, Filippo Maria Gambetta, Joel Klassen, Ashley Montanaro, Stephen Piddock, Raul A Santos, and Evan Sheridan. Towards near-term quantum simulation of materials.Nature Communications, 15(1):211, 2024. QUANTUM AGENTS 43

  12. [20]

    Simulating physics with computers, international journal of theoretical physics, 1982

    Richard P Feynman. Simulating physics with computers, international journal of theoretical physics, 1982

  13. [21]

    Cou- pled cluster method tailored with quantum computing.Physical Review Research, 6(2):023230, 2024

    Luca Erhart, Yuichiro Yoshida, Viktor Khinevich, and Wataru Mizukami. Cou- pled cluster method tailored with quantum computing.Physical Review Research, 6(2):023230, 2024

  14. [22]

    Yunheng Zou, Austin H. Cheng, Abdulrahman Aldossary, Jiaru Bai, Shi Xuan Leong, Jorge Arturo Campos-Gonzalez-Angulo, Changhyeok Choi, Cher Tian Ser, Gary Tom, Andrew Wang, Zijian Zhang, Ilya Yakavets, Han Hao, Chris Crebolder, Varinia Bernales, and Al´ an Aspuru-Guzik. El agen...

  15. [23]

    Department of Energy, Office of Science

    U.S. Department of Energy, Office of Science. Quantum information science applica- tions roadmap for the u.s. department of energy. Technical report, U.S. Department of Energy, Washington, DC, October 2024. Prepared by the DOE QIS Roadmap Committee, chaired by Jens Koch

  16. [24]

    Model context protocol (mcp): Landscape, security threats, and future research directions.arXiv preprint arXiv:2503.23278, 2025

    Xinyi Hou, Yanjie Zhao, Shenao Wang, and Haoyu Wang. Model context protocol (mcp): Landscape, security threats, and future research directions.arXiv preprint arXiv:2503.23278, 2025

  17. [25]

    Chaotic quantum encryption to secure image data in post quantum consumer tech- nology.IEEE Transactions on Consumer Electronics, 2024

    Muhammad Shahbaz Khan, Jawad Ahmad, Ahmed Al-Dubai, Nikolaos Pitropakis, Baraq Ghaleb, Amjad Ullah, Muhammad Attique Khan, and William J Buchanan. Chaotic quantum encryption to secure image data in post quantum consumer tech- nology.IEEE Transactions on Consumer Electronics, 2024

  18. [26]

    Generative ai meets responsible ai: Practical challenges and opportunities

    Krishnaram Kenthapadi, Himabindu Lakkaraju, and Nazneen Rajani. Generative ai meets responsible ai: Practical challenges and opportunities. InProceedings of the 29th ACM SIGKDD conference on knowledge discovery and data mining, pages 5805–5806, 2023

  19. [27]

    Quixer: A quan- tum transformer model.arXiv preprint arXiv:2406.04305, 2024

    Nikhil Khatri, Gabriel Matos, Luuk Coopmans, and Stephen Clark. Quixer: A quan- tum transformer model.arXiv preprint arXiv:2406.04305, 2024

  20. [28]

    A survey of quantum transformers: Approaches, ad- vantages, challenges, and future directions.arXiv preprint arXiv:2504.03192, 2025

    Hui Zhang and Qinglin Zhao. A survey of quantum transformers: Approaches, ad- vantages, challenges, and future directions.arXiv preprint arXiv:2504.03192, 2025

  21. [29]

    Quantum singular value transformation and beyond: exponential improvements for quantum matrix arithmetics

    Andr´ as Gily´ en, Yuan Su, Guang Hao Low, and Nathan Wiebe. Quantum singular value transformation and beyond: exponential improvements for quantum matrix arithmetics. InProceedings of the 51st annual ACM SIGACT symposium on theory of computing, pages 193–204, 2019

  22. [30]

    Walters, and Fei Wang

    Avin Seneviratne, Peter L. Walters, and Fei Wang. Polynomial time and space quan- tum algorithm for the simulation of non-markovian quantum dynamics, 2024

  23. [31]

    Single-shot quantum machine learning.Physical Review A, 111(4):042420, 2025

    Erik Recio-Armengol, Jens Eisert, and Johannes Jakob Meyer. Single-shot quantum machine learning.Physical Review A, 111(4):042420, 2025

  24. [32]

    Quantum reinforcement learning of classical rare dynamics: Enhancement by intrinsic fourier features.arXiv preprint arXiv:2504.16258, 2025

    Alissa Wilms, Laura Ohff, Andrea Skolik, Jens Eisert, Sumeet Khatri, and David A Reiss. Quantum reinforcement learning of classical rare dynamics: Enhancement by intrinsic fourier features.arXiv preprint arXiv:2504.16258, 2025

  25. [33]

    Towards provably efficient quantum algorithms for large- scale machine-learning models.Nature Communications, 15(1):434, 2024

    Junyu Liu, Minzhao Liu, Jin-Peng Liu, Ziyu Ye, Yunfei Wang, Yuri Alexeev, Jens Eisert, and Liang Jiang. Towards provably efficient quantum algorithms for large- scale machine-learning models.Nature Communications, 15(1):434, 2024. 44 SULTANOW, TEHRANI, DUTTA, BUCHANAN & KHAN

  26. [34]

    A survey of scaling in large language model reasoning, 2025

    Zihan Chen, Song Wang, Zhen Tan, Xingbo Fu, Zhenyu Lei, Peng Wang, Huan Liu, Cong Shen, and Jundong Li. A survey of scaling in large language model reasoning, 2025

  27. [35]

    Quantum amplitude amplification and estimation, 2002

    Gilles Brassard, Peter Høyer, Michele Mosca, and Alain Tapp. Quantum amplitude amplification and estimation, 2002

  28. [36]

    Quantum multiple kernel learning.arXiv preprint arXiv:2011.09694, 2020

    Seyed Shakib Vedaie, Moslem Noori, Jaspreet S Oberoi, Barry C Sanders, and Ehsan Zahedinejad. Quantum multiple kernel learning.arXiv preprint arXiv:2011.09694, 2020

  29. [37]

    Probabilistic quantum memories.Physical review letters, 87(6):067901, 2001

    Carlo A Trugenberger. Probabilistic quantum memories.Physical review letters, 87(6):067901, 2001

  30. [38]

    Quantum cryptography: Public key distri- bution and coin tossing.Theoretical computer science, 560:7–11, 2014

    Charles H Bennett and Gilles Brassard. Quantum cryptography: Public key distri- bution and coin tossing.Theoretical computer science, 560:7–11, 2014

  31. [39]

    Quantum cryptography using any two nonorthogonal states

    Charles H Bennett. Quantum cryptography using any two nonorthogonal states. Physical review letters, 68(21):3121, 1992

  32. [40]

    Quantum cryp- tography protocols robust against photon number splitting attacks¡? format?¿ for weak laser pulse implementations.Physical review letters, 92(5):057901, 2004

    Valerio Scarani, Antonio Acin, Gr´ egoire Ribordy, and Nicolas Gisin. Quantum cryp- tography protocols robust against photon number splitting attacks¡? format?¿ for weak laser pulse implementations.Physical review letters, 92(5):057901, 2004

  33. [41]

    Quantum cryptography based on bell’s theorem.Physical review letters, 67(6):661, 1991

    Artur K Ekert. Quantum cryptography based on bell’s theorem.Physical review letters, 67(6):661, 1991

  34. [42]

    Quantum cryptography without bell’s theorem.Physical review letters, 68(5):557, 1992

    Charles H Bennett, Gilles Brassard, and N David Mermin. Quantum cryptography without bell’s theorem.Physical review letters, 68(5):557, 1992

  35. [43]

    Authentication of quantum messages

    Howard Barnum, Claude Cr´ epeau, Daniel Gottesman, Adam Smith, and Alain Tapp. Authentication of quantum messages. InThe 43rd Annual IEEE Symposium on Foundations of Computer Science, 2002. Proceedings., pages 449–458. IEEE, 2002

  36. [44]

    Quantum digital signatures with quantum-key-distribution components.Physical Review A, 91(4):042304, 2015

    Petros Wallden, Vedran Dunjko, Adrian Kent, and Erika Andersson. Quantum digital signatures with quantum-key-distribution components.Physical Review A, 91(4):042304, 2015

  37. [45]

    Beating the fault-tolerance bound and security loopholes for byzantine agreement with a quantum solution.Research, 6:0272, 2023

    Chen-Xun Weng, Rui-Qi Gao, Yu Bao, Bing-Hong Li, Wen-Bo Liu, Yuan-Mei Xie, Yu-Shuo Lu, Hua-Lei Yin, and Zeng-Bing Chen. Beating the fault-tolerance bound and security loopholes for byzantine agreement with a quantum solution.Research, 6:0272, 2023. QUANTUM AGENTS 45 Eldar Sult...

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.