Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
156 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Out-of-distribution generalization for learning quantum dynamics (2204.10268v3)

Published 21 Apr 2022 in quant-ph, cs.LG, and stat.ML

Abstract: Generalization bounds are a critical tool to assess the training data requirements of Quantum Machine Learning (QML). Recent work has established guarantees for in-distribution generalization of quantum neural networks (QNNs), where training and testing data are drawn from the same data distribution. However, there are currently no results on out-of-distribution generalization in QML, where we require a trained model to perform well even on data drawn from a different distribution to the training distribution. Here, we prove out-of-distribution generalization for the task of learning an unknown unitary. In particular, we show that one can learn the action of a unitary on entangled states having trained only product states. Since product states can be prepared using only single-qubit gates, this advances the prospects of learning quantum dynamics on near term quantum hardware, and further opens up new methods for both the classical and quantum compilation of quantum circuits.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (51)
  1. M. Schuld, I. Sinayskiy, and F. Petruccione, An introduction to quantum machine learning, Contemporary Physics 56, 172 (2015).
  2. M. Schuld and F. Petruccione, Machine Learning with Quantum Computers (Springer, 2021).
  3. Y. Liu, S. Arunachalam, and K. Temme, A rigorous and robust quantum speed-up in supervised machine learning, Nature Physics , 1 (2021).
  4. D. Aharonov, J. Cotler, and X.-L. Qi, Quantum algorithmic measurement, Nature Communications 13, 1 (2022).
  5. H.-Y. Huang, R. Kueng, and J. Preskill, Information-theoretic bounds on quantum advantage in machine learning, Phys. Rev. Lett. 126, 190505 (2021a).
  6. J. Cotler, H.-Y. Huang, and J. R. McClean, Revisiting dequantization and quantum advantage in learning tasks, arXiv preprint arXiv:2112.00811  (2021).
  7. M. C. Caro and I. Datta, Pseudo-dimension of quantum circuits, Quantum Machine Intelligence 2, 14 (2020).
  8. L. Banchi, J. Pereira, and S. Pirandola, Generalization in quantum machine learning: A quantum information standpoint, PRX Quantum 2, 040321 (2021).
  9. C. M. Popescu, Learning bounds for quantum circuits in the agnostic setting, Quantum Information Processing 20, 1 (2021).
  10. H. Cai, Q. Ye, and D.-L. Deng, Sample complexity of learning parametric quantum circuits, Quantum Science and Technology 7, 025014 (2022).
  11. T. Volkoff, Z. Holmes, and A. Sornborger, Universal compiling and (no-)free-lunch theorems for continuous-variable quantum learning, PRX Quantum 2, 040327 (2021).
  12. J. Preskill, Quantum computing in the NISQ era and beyond, Quantum 2, 79 (2018).
  13. C. L. Degen, F. Reinhard, and P. Cappellaro, Quantum sensing, Rev. Mod. Phys. 89, 035002 (2017).
  14. T. Jones and S. C. Benjamin, Robust quantum compilation and circuit optimisation via energy minimisation, Quantum 6, 628 (2022).
  15. H.-Y. Hu, S. Choi, and Y.-Z. You, Classical shadow tomography with locally scrambled quantum dynamics, Physical Review Research 5, 023027 (2023).
  16. V. N. Vapnik and A. Y. Chervonenkis, On the uniform convergence of relative frequencies of events to their probabilities, Th. Prob. App. 16, 264 (1971).
  17. L. G. Valiant, A theory of the learnable, Communications of the ACM 27, 1134 (1984).
  18. D. Gottesman and I. Chuang, Quantum digital signatures, arXiv preprint quant-ph/0105032  (2001).
  19. H. Shimodaira, Improving predictive inference under covariate shift by weighting the log-likelihood function, Journal of statistical planning and inference 90, 227 (2000).
  20. S. J. Pan and Q. Yang, A survey on transfer learning, IEEE Transactions on knowledge and data engineering 22, 1345 (2009).
  21. Note, we use the terms ‘distribution’ and ‘ensemble’ interchangeably.
  22. Here and elsewhere, the “∼similar-to\sim∼” notation means that the random variable on the left has the distribution on the right as its law. For instance, U∼𝒰LSsimilar-to𝑈subscript𝒰LSU\sim\mathcal{U}_{\rm LS}italic_U ∼ caligraphic_U start_POSTSUBSCRIPT roman_LS end_POSTSUBSCRIPT means that the random unitary U𝑈Uitalic_U is drawn from the distribution 𝒰LSsubscript𝒰LS\mathcal{U}_{\rm LS}caligraphic_U start_POSTSUBSCRIPT roman_LS end_POSTSUBSCRIPT.
  23. Here and elsewhere, 𝒰⁢|0⟩⊗n𝒰superscriptket0tensor-productabsent𝑛\mathcal{U}|0\rangle^{\otimes n}caligraphic_U | 0 ⟩ start_POSTSUPERSCRIPT ⊗ italic_n end_POSTSUPERSCRIPT denotes the ensemble of states generated by drawing unitaries from 𝒰𝒰\mathcal{U}caligraphic_U and applying them to the n𝑛nitalic_n-qubit all-zero state |0⟩⊗nsuperscriptket0tensor-productabsent𝑛|0\rangle^{\otimes n}| 0 ⟩ start_POSTSUPERSCRIPT ⊗ italic_n end_POSTSUPERSCRIPT.
  24. D. Stilck França and R. Garcia-Patron, Limitations of optimization algorithms on noisy quantum devices, Nature Physics 17, 1221 (2021).
  25. H.-Y. Huang, R. Kueng, and J. Preskill, Predicting many properties of a quantum system from very few measurements, Nature Physics 16, 1050 (2020).
  26. J. A. Nelder and R. Mead, A simplex method for function minimization, The Computer Journal 7, 308 (1965).
  27. E. Knill and R. Laflamme, Power of one bit of quantum information, Physical Review Letters 81, 5672 (1998).
  28. D. A. Roberts and B. Yoshida, Chaos and complexity by design, Journal of High Energy Physics 2017, 121 (2017).
  29. S. T. Flammia and J. J. Wallman, Efficient estimation of pauli channels, ACM Transactions on Quantum Computing 1, 1 (2020).
  30. R. Harper, S. T. Flammia, and J. J. Wallman, Efficient learning of quantum noise, Nature Physics 16, 1184 (2020).
  31. R. Harper, W. Yu, and S. T. Flammia, Fast estimation of sparse quantum noise, PRX Quantum 2, 010322 (2021).
  32. S. T. Flammia and R. O’Donnell, Pauli error estimation via population recovery, Quantum 5, 549 (2021).
  33. M. C. Caro, Binary classification with classical instances and quantum labels, Quantum Machine Intelligence 3, 18 (2021).
  34. M. Fanizza, Y. Quek, and M. Rosati, Learning quantum processes without input control, arXiv preprint arXiv:2211.05005  (2022).
  35. H.-Y. Huang, S. T. Flammia, and J. Preskill, Foundations for learning from noisy quantum experiments, arXiv preprint arXiv:2204.13691  (2022b).
  36. H.-Y. Huang, S. Chen, and J. Preskill, Learning to predict arbitrary quantum processes, arXiv preprint arXiv:2210.14894  (2022c).
  37. M. C. Caro, Learning quantum processes and Hamiltonians via the Pauli transfer matrix, arXiv preprint arXiv:2212.04471  (2022).
  38. A. Uvarov, A. Kardashin, and J. D. Biamonte, Machine learning phase transitions with a quantum processor, Physical Review A 102, 012415 (2020).
  39. M. Arjovsky, Out of distribution generalization in machine learning, arXiv preprint arXiv:2103.02667  (2021).
  40. M. A. Nielsen, A simple formula for the average gate fidelity of a quantum dynamical operation, Physics Letters A 303, 249 (2002).
  41. C. B. Mendl and M. M. Wolf, Unital quantum channels–convex structure and revivals of Birkhoff’s theorem, Communications in Mathematical Physics 289, 1057 (2009).
  42. F. G. Brandao, A. W. Harrow, and M. Horodecki, Local random quantum circuits are approximate polynomial-designs, Communications in Mathematical Physics 346, 397 (2016).
  43. A. Harrow and S. Mehraban, Approximate unitary t𝑡titalic_t-designs by short random quantum circuits using nearest-neighbor and long-range gates, arXiv preprint arXiv:1809.06957  (2018).
  44. J. Haferkamp, Random quantum circuits are approximate unitary t𝑡titalic_t-designs in depth O⁢(n⁢t5+o⁢(1))𝑂𝑛superscript𝑡5𝑜1O\left(nt^{5+o(1)}\right)italic_O ( italic_n italic_t start_POSTSUPERSCRIPT 5 + italic_o ( 1 ) end_POSTSUPERSCRIPT ), Quantum 6, 795 (2022).
  45. D. Gross, K. Audenaert, and J. Eisert, Evenly distributed unitaries: On the structure of unitary designs, Journal of mathematical physics 48, 052104 (2007).
  46. R. Kueng and D. Gross, Qubit stabilizer states are complex projective 3-designs, arXiv preprint arXiv:1510.02767  (2015).
  47. Z. Webb, The clifford group forms a unitary 3-design, Quantum Information and Computation 16, 1379 (2016).
  48. H. Zhu, Multiqubit clifford groups are unitary 3-designs, Physical Review A 96, 062336 (2017).
  49. F. Vatan and C. Williams, Optimal quantum circuits for general two-qubit gates, Physical Review A 69, 032315 (2004).
  50. V. V. Shende, I. L. Markov, and S. S. Bullock, Minimal universal two-qubit controlled-not-based circuits, Physical Review A 69, 062321 (2004).
  51. R. R. Tucci, An Introduction to Cartan’s KAK Decomposition for QC Programmers, arXiv:quant-ph/0507171  (2005).
Citations (72)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now