noa/thesis
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

text improvements

Browse source
This commit is contained in:
Noa Aarts 2026-01-06 11:20:07 +01:00
parent 428cfd58b7
commit 14bb387e2d
Signed by: noa
GPG key ID: 1850932741EFF672
1 changed files with 7 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

14
planning/plan.typ
View file

@ -1,11 +1,13 @@
== Introduction
What I'll be making is a program that uses a Hardware discriptor
= Introduction
The goal of this project is to make a program that uses a Hardware discriptor
-like the one exported by qiskit- together with a target expressibility,
target entanglement, and a minimal Fidelity to produce a
diverse set of Parametrised Quantum Circuits (PQCs).
set of Parametrised Quantum Circuits (PQCs) that cover different parts of the Hilbert Space.
Unlike previous training free Quantum Architecture Searches@training-free@evolutionary-architecture-search
we will make the search hardware-specific while also keeping it training-free
and target a specific expressibility instead of just maximizing to help mitigate barren plateaus.
we will make the search hardware-specific, by implementing noise proxies,
and target a specific expressibility instead of maximizing it to help mitigate barren plateaus,
another differentiator is the optimisation for diversity which should increase the probability
that a good PQC for a problem is in the produced set of PQCs.
The program will be using one of the Quality-Diversity Evolutionary Algorithms@qdea to achieve this,
the quality aspect will use the Expressibility, Entanglement, and simulated Noise -or proxies for these-
to determine the Quality, while some measure of distance between different expressible Hilbert Spaces
@ -35,6 +37,4 @@ In the next parts I'll list the inputs and outputs in a more structured way.
- Fidelity
== References
#bibliography("references.bib")