Finance is a complex area with various formulas, statistics and models. But that mathematical language could be amenable to using a quantum computer for part of the processes that finance professionals typing undertake. A new book by Anshul Saxena, Javier Mancilla, Iraitz Montalban and Christophe Pere bridges the gap between traditional algorithms for finance and the using a quantum algorithm.
Whether portfolio design or options, researchers have been exploring how quantum algorithms (a quantum way) could impact these processes. Still, as always with new areas, getting started can be particularly difficult, and that is where resources like “Financial Modelling using Quantum Computing” become very useful because they provide that all-important recipe book of code snippets (python and mainly qiskit) and examples that enable readers to bootstrap ideas quickly.
The book does also cover working with D-wave and the quantum annealer, which is used for specific optimization type problems and Xanadu’s Pennylane – which is an excellent resource for people looking at QML or Quantum Machine Learning as it has made some fantastic tools to make QML very straight-forward.
The book assumes a rough working knowledge of Python and qiskit, and that is a fairly safe assumption for these days. The go-to language of choice for many projects is Python. The most popular quantum framework is qiskit, so you can be sure that the examples impact the most widest addressable market of readers. We loved that the book went to work; there was no initial yawn at five chapters of introduction to quantum gate filler materials. What was great to see the all-important reference to the complexity of algorithms – the whole point of invoking quantum in the first place! If anyone wants an introduction to quantum computing, maths and gates, plenty of great books and learning resources are out there. The book references the quantum ecosystem and other frameworks and languages and introduces a couple of popular algorithms such as Josza, Grover and Shor (although the detail is not covered).
Quantum Machine Learning: QSVC
For example, many techniques rely on QML or Quantum Machine Learning techniques in the Quantum Equivalent of Support Vector Machines or QSVC. Such methods can be helpful for classification – as the name suggests. The context illustrated is in credit risk, where the aim is to classify data and learn the distinction between (often two) classes. Using features of Qiskit, the user is then walked through the process of encoding classical data and performing the quantum equivalent of the classical SVM – the QSVC. The format is much the same for other problems, doing a side-by-side walkthrough of the classical process and then illustrating how a quantum process would work.
Encoding: Classical -> Quantum
There are multiple methods of encoding classical data into quantum representation, and the book will feature how to do this. Of course, the heavy lifting can be done with the use of the inbuilt Qiskit libraries, but what is crucial is that this book isn’t going to give you every technique – a whole book itself. However, you will get a motivation for encoding, which is a non-trivial problem. The point of the book is clear, it’s to bootstrap your understanding to be able to test this out without mountains of extraneous information clouding the picture.
QNNs: Quantum Neural Networks
Neural networks and deep learning have garnered much interest of the past few years with the likes of Deep Neural Networks, and now LLM’s are Large Language Models being widespread and fairly common. It should be no surprise that there are quantum equivalents of NNs named QNN’s. Again a subset of QML, they can be used in many processes, and the book highlights their use in Credit Risk Analytics.
A multitude of Quantum Machine Learning techniques
You’ll also be introduced to VQC (Variational Quantum Circuits) using Xanadu’s PennyLane toolkit. Variational techniques allow a quantum circuit to be parameterized and thus effectively learn a given target output.
As you’d expect from any finance book worth its salt, a chapter on stock portfolio optimization uses the D-wave quantum annealer (a different type of technology from gate-based quantum computing). Then there is derivatives pricing for those that like their options quantum!
Aside from code snippets, there are plenty of quantum clouds and even some great insight into the quantum landscape, including the myriad of clouds and platforms.
Whilst it’s hard to shy away from the dominance of Qiskit frameworks, it’s been refreshing to see D-wave and PennyLane in the mix here, and we applaud that.
Quick Summary
For some, it might seem a little light on theory. But given the author’s numerate and technical background, this is intentional. It’s a quick-start guide and should be viewed as that, and it does as it says; it provides much-needed context and practical application, often introduced from an overly technical and theoretical base. This gets the reader to the key concepts and techniques without tying up the reader in huge amounts of theory and unnecessary complication. A competent finance professional could pick this up, we think, and make some progress. Sure, they won’t become an expert on all things quantum, but they will have that “jumping off” point.
Table of Contents
You can get a copy from Amazon or other good bookseller.
