Seminario: «Adiabatic training for Variational Quantum Algorithms»

Seminario impartido por Ernesto Acosta. PhD Student UGR.

Abstract: 

On this talk we present a new hybrid Quantum Machine Learning model (QML) composed of three elements: a classical computer in charge of the data preparation and interpretation; a Gate-based Quantum Computer running the Variational Quantum Algorithm (VQA) representing a Quantum Neural Network (QNN); and an adiabatic Quantum Computer where the optimization function is executed to find the best parameters for the VQA.

As of the present moment the majority of VQAs are being trained using gradient-based classical optimizers having to deal with the barren-plateau effect. Some gradient-free classical approaches such as Evolutionary Algorithms have also been proposed to overcome this effect. However, adiabatic quantum models have not been defined to train VQAs.

A quick review of Artificial Neural Networks and Variational Quantum Algorithms concepts is presented, along with the description of barren-plateau effect in relation to vanishing gradients. Then we will describe the proposed adiabatic training model comparing the obtained results against the classical gradient-based algorithms, showing the feasibility of integration for gate-based and adiabatic quantum computing models.

We will end up with some highlights on the current phase of research towards iterative and multithreaded adiabatic training.

• Fecha: Viernes, 25 de octubre de 2024
• Hora: 11:00 h.
• Lugar: Facultad de Ciencias. Laboratorio de Física Computacional (Edificio de Física, planta baja, junto al péndulo).
• Organiza: Quantum Thermodynamics and Quantum Computing Group (QTCG)