Quantum
Machine Learning Conference 2023

Date: 18.11.2023
Form: Online event
Recording

Explore previous editions
2021 | 2022

Agenda (times in UTC+1)

9:30
Mandaar Pande, "Overview of Mathematical Foundations for Quantum Computing and Quantum Machine Learning".
10:00
Tomasz Rybotycki, "Training Quantum Neural Networks with Metaheuristics".
10:30
Mateusz Ostaszewski, "Quantum Architecture Search".
11:00
Manish Modani, "CUDA-Quantum: The platform for Integrated & Accelerated Quantum Classical Computing"
11:30
Amlan Chakrabarti, "Unlocking the Quantum Frontier: Exploring the Fascinating World of Quantum Machine Learning".
12:00
Iraitz Montalban, "Challenges for industry relevant Quantum Computing".

Speakers

BIO: After my Ph.D. in Theoretical Quantum Physics from Hyderabad Central University in 1994, I come with more than 29 years of total experience with over 9 years in Academics, and around 20 years in the IT Industry at various technical, consultancy and managerial positions. Currently, I am researching in the areas of Quantum Computing, specifically Quantum Machine Learning, Quantum Communication, and Quantum Internet.

Mandaar Pande

Overview of Mathematical Foundations
for Quantum Computing and Quantum Machine Learning

Abstract: The talk will give an overview of the mathematics required to start off on the journey of quantum computing and quantum machine learning. It will cover the mathematical perspectives of the three basic principles of quantum mechanics namely superposition, entanglement, and interference which are the basis for quantum computing. The talk will also give an overview of the fundamental steps involved in a QML problem focussing on the various types of data encoding techniques required in converting data into quantum states.

BIO: Dr Tomasz Rybotycki has obtained his Ph.D. in 2023 from the Systems Research Institute of the Polish Academy of Sciences, Warsaw, Poland. His doctoral advisor was prof. Piotr Kulczycki.

During his Ph.D., dr Rybotycki worked on the predictive estimation of the density of non-stationary data streams, where he designed a kernel density estimator-based algorithm for predictive estimation of data density. The title of his thesis was “Estimation of data density for non-stationary streaming data”. During his Ph.D. he also obtained a B.Sc. degree in Physics, after successfully defending his B.Sc. thesis entitled “Framework for performing experiments on IBM Quantum Computers”.

Since 2017, dr Rybotycki works in the Systems Research Institute of the Polish Academy of Sciences as a research assistant.

In 2020, dr Rybotycki obtained Research Scholarship at AstroCeNT, where he investigated the use of metaheuristics for quantum neural networks training. The same year he started working at Centre for Theoretical Physics of the Polish Academy of Sciences, initially as a researcher / software engineer, and later as a (theoretical physics) Ph.D. student in a project. He left the Centre for Theoretical Physics in 2022.

In 2022, dr Rybotycki started working at ACK Cyfronet AGH, where, as a senior programmer, he is responsible for preparing an auto quantum machine learning e-platform.

Dr Rybotycki joined AstroCeNT again in 2023. Currently he works on Quantum Machine Learning with prof. Piotr Gawron. His research focuses on quantum circuits optimization using ZX calculus and applying (Q)ML techniques for earth observation.

Tomasz Rybotycki

Training Quantum Neural Networks with Metaheuristics

Abstract: This study investigates the viability of metaheuristic optimization techniques to train quantum neural networks (QNNs) while assessing their feasibility, accuracy, and resource requirements. As QNNs hold the potential to tackle complex computational challenges beyond classical neural networks' capabilities, their training remains a daunting task due to exponential growth in parameter space and their resource-intensive nature.

BIO: Mateusz Ostaszewski earned his Ph.D. with distinction from the Institute of Theoretical and Applied Informatics, Polish Academy of Sciences in Gliwice. While preparing his doctoral thesis, he conducted a research internship at University College London (UCL) in the Computer Science Lab - Quantum Information. Following his doctoral studies, he initiated collaboration with the research group led by Prof. Vedran Dunjko at Leiden University. He is undertaking a postdoctoral fellowship at the Warsaw University of Technology.

He has been a recipient of the Ministry of Science and Higher Education scholarship for outstanding achievements, as well as an Etiuda doctoral scholarship. Moreover, he actively contributed to research projects funded by the National Science Centre (NCN) and the Foundation for Polish Science (FNP). His research findings have been presented at conferences such as NeurIPS and CoLLAs and published in Quantum, Computer Physics Communications, and Journal of Physics A: Mathematical and Theoretical. His research interests are primarily centered around continual and reinforcement learning methodologies. In parallel, he is leveraging deep learning techniques for quantum circuit construction, particularly for noisy intermediate-scale quantum quantum devices.

Mateusz Ostaszewski

Quantum Architecture Search

Abstract: The paramount challenge in the era of noisy intermediate-scale quantum computing is the discovery of efficient quantum circuits that can operate effectively within the constraints of current quantum devices. Variational quantum algorithms (VQAs) offer a promising solution wherein the circuit architecture is first established and subsequently fine-tuned through parameter optimization to address specific quantum tasks. However, the optimization process can often become intractable, with overall performance heavily reliant on the initially chosen circuit architecture. To address these challenges, several quantum architecture search (QAS) algorithms have been developed to automate the selection of optimal circuit architectures. In this tutorial, I will overview the key methodological trends within Quantum Architecture Search. We will delve into the technical intricacies of various approaches, exploring their strengths and limitations. Additionally, I will shed light on the ongoing challenges that continue to engage researchers in this rapidly evolving field.

BIO: Dr Manish Modani has 18+ years of industry and worked in High Performance Computing (HPC), Artificial Intelligence, Quantum Computing. He has ported, benchmarked, and optimized HPC & HPC+AI applications related to Weather, Computational Fluid Dynamics, Insurance, Molecular Chemistry, Genomics etc. He has hands-on experience while working on various HPC architectures including Cray, IBM, CDAC PARAM series, including AIRAWAT PSAI system which is globally ranked at 75th. Manish work is patented and published in various peer reviewed journals. Recently, he has demonstrated the acceleration of popular quantum computing algorithm (eg QFT, Sycamore, Shor). Prior to joining NVIDIA, Manish was the Technical Lead for India/SA at IBM. During his PhD from Indian Institute of Technology (IIT) Delhi, India, Manish has developed air pollution models for the dispersion of air pollutants in low wind conditions.

Manish Modani

CUDA-Quantum:
The platform for Integrated & Accelerated
Quantum Classical Computing

Abstract: A critical challenge in quantum computers making is effectively combining them with classical computing resources. From the classical side of hybrid algorithms and integrated application workflows to decoding syndromes for quantum error correction, tightly coupled high performance classical computing will be important for many of the functions required to realize useful quantum computing (including Quantum Machine Learning). A key tool for enabling research and application development is a programming model and software toolchain which allow researchers to straightforwardly co-program classical and quantum computers and leverage the best tools available for each. The NVIDIA CUDA-Quantum is a single-source programming model in C++ and Python for heterogeneous quantum-classical computing. The QODA platform provides several advantages and new capabilities that enable users to get more out of quantum processors.  Here, CUDA-Quantum is discussed and demonstrated for several use cases including Variational Quantum Eigen solver (VQE) where it provides a significant (~287x) performance and capability benefit over existing quantum programming.

BIO: Dr. Amlan Chakrabarti, a Full Professor at the University of Calcutta’s A.K. Choudhury School of Information Technology, boasts over two decades of expertise in engineering education and research. With a Post-Doctoral fellowship from Princeton University (2011-2012), he has garnered numerous accolades, including the DST BOYSCAST fellowship, INSA Visiting Faculty Fellowship, JSPS Invitation Research Award, Erasmus Mundus Leaders Award, Hamied Visiting Professorship, Siksha Ratna Award, and Fellowship from the West Bengal Academy of Science and Technology. Dr. Chakrabarti leads the ALICE-India Collaboration at CERN and holds significant administrative roles in higher education and IT in West Bengal. His vast research, spanning 200+ papers and 20 Ph.D. students, focuses on machine learning, computer vision, and quantum computing.

Amlan Chakrabarti

Unlocking the Quantum Frontier: Exploring the Fascinating World of Quantum Machine Learning

Abstract: Quantum Machine Learning (QML) stands at the forefront of cutting-edge research, bridging the realms of quantum physics and artificial intelligence. In this captivating lecture, we embark on a journey to Unlock the Quantum Frontier, delving into the intricate interplay between quantum computing and machine learning. We will explore the fascinating potential of harnessing quantum principles to solve complex problems, exponentially accelerate computations, and revolutionize the world of data analysis. The lecture will uncover the promises and challenges of Quantum Machine Learning and glimpse the future of computing and AI, where quantum states hold the key to unlocking new horizons in technology and science.

BIO: Iraitz Montalban is a PhD candidate at the University of Basque Country holding masters in quantum computing, mathematical modelling and data protection. Co-author and lecturer in different institutions has held roles of responsibility for a world-wide utility and helped build the technology roadmap for many other medium and large size companies as a consultant. Currently leads the Product development at Kipu Quantum GmbH, a German start-up looking for near-term quantum advantage by leveraging application- and hardware-specific applications.

Iraitz Montalban

Challenges for industry relevant Quantum Computing

Abstract: Companies in all industry verticals face daily challenges on computing and simulating scenarios to improve their decision making process. From financial institutions to large logistic firms, they all face the complexities on how to efficiently leverage their resources, from all the potential combinations there might exist. These challenges are often framed within the Operational Research domain and large computing resources are needed to explore the vast solution space to select the best decision to be made by Business units. But Quantum Computing is aiming to revolutionize this decision making ability as well. The main issue is that this theoretical gain is harmed by many factors quantum computing needs to overcome in order to become such a disruptor. We will briefly explore some of those business use cases, showing how existing companies are actively involved in pushing this research field and how the challenges faced in currently available NISQ devices have shaped the techniques and algorithms landscape for Quantum Computers. The talk will cover the usage of QC for Industrial ML steps (ex. Quantum Feature Selection) but also how variational circuits are used as classifiers and preprocessing steps (ex. Quantum Kernels and QSVC).