Quantum Physics
[Submitted on 6 Oct 2022 (this version), latest version 27 Aug 2024 (v4)]
Title:A perturbative gadget for delaying the onset of barren plateaus in variational quantum algorithms
View PDFAbstract:Variational quantum algorithms are being explored as a promising approach to finding useful applications for noisy intermediate-scale quantum computers. However, cost functions corresponding to many problems of interest are inherently global, defined by Hamiltonians with many-body interactions. Consequently, the optimization landscape can exhibit exponentially vanishing gradients, so-called barren plateaus, rendering optimal solutions difficult to find. Strategies for mitigating barren plateaus are therefore needed to make variational quantum algorithms trainable and capable of running on larger-scale quantum devices. In this work, we contribute the toolbox of perturbative gadgets to the portfolio of methods being explored in the quest for making noisy intermediate-scale quantum devices useful. Specifically, we introduce a novel perturbative gadget, tailored to variational quantum algorithms, that can be used to avoid barren plateaus. Our perturbative gadget encodes an arbitrary many-body Hamiltonian corresponding to a global cost function into the low-energy subspace of a three-body Hamiltonian. Our construction requires $rk$ additional qubits for a $k$-body Hamiltonian comprising $r$ terms. We provide rigorous guarantees on the optimization of the local cost function defined by our three-body gadget Hamiltonian with respect to the original cost function, and we prove that this local cost function exhibits non-vanishing gradients, thus delaying the onset of barren plateaus. We then provide numerical demonstrations to show the functioning of our approach.
Submission history
From: Paul K. Faehrmann [view email][v1] Thu, 6 Oct 2022 17:51:23 UTC (2,257 KB)
[v2] Fri, 4 Nov 2022 14:40:41 UTC (2,261 KB)
[v3] Fri, 15 Sep 2023 13:41:33 UTC (2,542 KB)
[v4] Tue, 27 Aug 2024 13:42:07 UTC (2,760 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.