Information-theoretic generalization bounds for learning from quantum data

Matthias C. Caro, Tom Gur, Cambyse Rouzé, Daniel Stilck França, Sathyawageeswar Subramanian
Proceedings of Thirty Seventh Conference on Learning Theory, PMLR 247:775-839, 2024.

Abstract

Learning tasks play an increasingly prominent role in quantum information and computation. They range from fundamental problems such as state discrimination and metrology over the framework of quantum probably approximately correct (PAC) learning, to the recently proposed shadow variants of state tomography. However, the many directions of quantum learning theory have so far evolved separately. We propose a mathematical formalism for describing quantum learning by training on classical-quantum data and then testing how well the learned hypothesis generalizes to new data. In this framework, we prove bounds on the expected generalization error of a quantum learner in terms of classical and quantum information-theoretic quantities measuring how strongly the learner’s hypothesis depends on the data seen during training. To achieve this, we use tools from quantum optimal transport and quantum concentration inequalities to establish non-commutative versions of decoupling lemmas that underlie classical information-theoretic generalization bounds. Our framework encompasses and gives intuitive generalization bounds for a variety of quantum learning scenarios such as quantum state discrimination, PAC learning quantum states, quantum parameter estimation, and quantumly PAC learning classical functions. Thereby, our work lays a foundation for a unifying quantum information-theoretic perspective on quantum learning.

Cite this Paper

BibTeX
@InProceedings{pmlr-v247-caro24a, title = {Information-theoretic generalization bounds for learning from quantum data}, author = {Caro, Matthias C. and Gur, Tom and Rouz{\'e}, Cambyse and Stilck Fran\c{c}a, Daniel and Subramanian, Sathyawageeswar}, booktitle = {Proceedings of Thirty Seventh Conference on Learning Theory}, pages = {775--839}, year = {2024}, editor = {Agrawal, Shipra and Roth, Aaron}, volume = {247}, series = {Proceedings of Machine Learning Research}, month = {30 Jun--03 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v247/caro24a/caro24a.pdf}, url = {https://proceedings.mlr.press/v247/caro24a.html}, abstract = {Learning tasks play an increasingly prominent role in quantum information and computation. They range from fundamental problems such as state discrimination and metrology over the framework of quantum probably approximately correct (PAC) learning, to the recently proposed shadow variants of state tomography. However, the many directions of quantum learning theory have so far evolved separately. We propose a mathematical formalism for describing quantum learning by training on classical-quantum data and then testing how well the learned hypothesis generalizes to new data. In this framework, we prove bounds on the expected generalization error of a quantum learner in terms of classical and quantum information-theoretic quantities measuring how strongly the learner’s hypothesis depends on the data seen during training. To achieve this, we use tools from quantum optimal transport and quantum concentration inequalities to establish non-commutative versions of decoupling lemmas that underlie classical information-theoretic generalization bounds. Our framework encompasses and gives intuitive generalization bounds for a variety of quantum learning scenarios such as quantum state discrimination, PAC learning quantum states, quantum parameter estimation, and quantumly PAC learning classical functions. Thereby, our work lays a foundation for a unifying quantum information-theoretic perspective on quantum learning.} }
Endnote
%0 Conference Paper %T Information-theoretic generalization bounds for learning from quantum data %A Matthias C. Caro %A Tom Gur %A Cambyse Rouzé %A Daniel Stilck França %A Sathyawageeswar Subramanian %B Proceedings of Thirty Seventh Conference on Learning Theory %C Proceedings of Machine Learning Research %D 2024 %E Shipra Agrawal %E Aaron Roth %F pmlr-v247-caro24a %I PMLR %P 775--839 %U https://proceedings.mlr.press/v247/caro24a.html %V 247 %X Learning tasks play an increasingly prominent role in quantum information and computation. They range from fundamental problems such as state discrimination and metrology over the framework of quantum probably approximately correct (PAC) learning, to the recently proposed shadow variants of state tomography. However, the many directions of quantum learning theory have so far evolved separately. We propose a mathematical formalism for describing quantum learning by training on classical-quantum data and then testing how well the learned hypothesis generalizes to new data. In this framework, we prove bounds on the expected generalization error of a quantum learner in terms of classical and quantum information-theoretic quantities measuring how strongly the learner’s hypothesis depends on the data seen during training. To achieve this, we use tools from quantum optimal transport and quantum concentration inequalities to establish non-commutative versions of decoupling lemmas that underlie classical information-theoretic generalization bounds. Our framework encompasses and gives intuitive generalization bounds for a variety of quantum learning scenarios such as quantum state discrimination, PAC learning quantum states, quantum parameter estimation, and quantumly PAC learning classical functions. Thereby, our work lays a foundation for a unifying quantum information-theoretic perspective on quantum learning.
APA
Caro, M.C., Gur, T., Rouzé, C., Stilck França, D. & Subramanian, S.. (2024). Information-theoretic generalization bounds for learning from quantum data. Proceedings of Thirty Seventh Conference on Learning Theory, in Proceedings of Machine Learning Research 247:775-839 Available from https://proceedings.mlr.press/v247/caro24a.html.

Related Material