# Why Quantum Artificial Intelligence with Qiskit

Here's why Andreas Wichert is publishing a book in early 2024 covering quantum artificial intelligence using IBM's Qiskit.

Steve Kelly

9/2/20236 min read

Professor Andreas (Andrzej) Wichert is publishing a book called "__Quantum Artificial Intelligence with Qiskit__" in January 2024, aimed at providing a coherent overview of the Quantum Artificial Intelligence (QAI) field and providing the necessary resources for his readers to create and use quantum programs on their everyday hardware. The book takes a look at symbolical quantum algorithms, quantum Machine Learning, and gives the reader a deeper understanding of how these algorithms and concepts can and will be used now and in the near future.

His target audience with the book is a general everyday user, along with students and instructors in courses related to computer science, artificial intelligence, data science, and machine learning.

The book is slated to be 334 pages with 135 black and white illustrations. It will be published by CRC Press/Taylor & Francis.

So why should anyone be concerned with giving a look at Quantum Artificial Intelligence using cloud quantum computing provided by IBM's Qiskit platform? To simply say why you should be interested is twofold:

1) You have certainly heard and probably used artificial intelligence.

2) You have probably heard about quantum computing and how it provides faster processing speeds than traditional computing.

Combine __quantum computing and machine learning and you get very fast processing speeds__ that as of this time require powerful and expensive hardware which the average consumer cannot afford or easily acquire. Granted, this is if you want to locally train and run a large language model on your own hardware. Of course, we know about Chat-GPT and the various other flavors of large language models, but nevertheless Dr. Andrzej Wichert is going to show how this field is growing and expanding to average everyday users.

Before we go a bit deeper into quantum artificial intelligence, let's take a look at who Professor Andreas (Andrezej) Wichert is.

__Dr. Andreas (Andrezej) Wichert__ is a professor from the Universidade de Lisboa (Technical University of Lisbon - Department of Computer Science and Engineering) and is a researcher with the INESC-ID at the same institution. He received a PHD in computer science in 2000 with his work on associative computing. His research is focused on artificial intelligence, machine learning, neural networks, quantum cognition, and quantum artificial intelligence. He has on __his Google Scholar__ 120 articles as a co-author and an author. His articles have been cited 1477 times all-time, 819 times since 2018. He has an overall h-index of 21, since 2018 his h-index is 15. His i10-index all time is 41, and since 2018 is 19. His top three cited papers according to Google Scholar are: "From symbolic to sub-symbolic information in question classification" published in 2011 with 232 citations; "Principles of quantum artificial intelligence: quantum problem solving and machine learning" published in 2020 with 79 citations; and "Quantum-like Bayesian networks for modeling decision making" published in 2016 with 71 citations. He was born in Wroclaw, Poland.

According to __his ResearchGate profile__, his top co-authors are V. Moret-Bonillo, Prayag Tiwari, Gerhard Glatting, Luisa Coheur, and Luis Sacouto. His latest books are: "__Mind, Brain, Quantum Ai, and the Multiverse__", "Machine Learning- A Journey to Deep Learning with Exercises and Answers", and "Principles of Quantum Artificial Intelligence: Quantum Problem Solving and Machine Learning, 2nd Edition".

Needless to say, Dr. Andreas (Andrezej) Wichert is a technical expert in the field of machine learning, quantum artificial intelligence, and is well-respected in his fields of study. We can expect a well-researched and thorough exploration of QAI with Qiskit when his book is released in January 2024.

Ok, let's go a bit into quantum artificial intelligence and why we should take note of it. Dr. Andreas Wichert talks about __an example of how artificial intelligence relates to quantum computing__ in his book Principles of Quantum Artificial Intelligence: Quantum Problem Solving and Machine Learning: "the artificial intelligence framework, such as search and production system theory, allows an elegant description of a quantum computer model that is capable of quickly executing programs." Okay, quantum computing can make artificial intelligence run faster. How much faster can quantum computing make artificial intelligence algorithms run compared to the traditional computing?

While I was able to find a specific percentage of quantum computing versus traditional computing models for machine learning/artificial intelligence problems, which is said to be about 63% faster than a classical computer according to Valeria Saggio, a quantum physicist at Massachusetts Institute of Technology. This was not fully satisfying, and I was able to stumble upon a couple of papers by Google and IBM that show (rather technically) how the improvements happen. Both of the papers published talk about the use of kernels which is a measure showing the relatedness of data points in respect to particular features. The example given from the website well-respected Quanta Magazine in their article "__Machine Learning Gets a Quantum Speedup__" uses three colors, Red, Blue, and Orange. Red and Orange are neighbors when looking at them as colors. However, when you compare them in their number of characters with blue having four, Red having three, and Orange having six, Blue sits between Red and Orange. They state that "kernels are like lenses that allow an algorithm to classify data in different ways to find patterns that help distinguish future inputs". To state simply: kernels help algorithms find new ways of looking at data.

The relationship of kernels to quantum computing is that they act similarly when doing data manipulation. We could think of this as clustering data in different respects to produce outputs which predict future inputs, and thus future outputs and so on. This is how neural networks work in both the unsupervised and supervised case. Neural networks are the heart of artificial intelligence. These researchers have applied quantum computing to kernels calling them quantum kernels, adding quantum superpositions to the mix to classify data in relation to each other, with the data appearing random initially but over time the data starts to take on a life of its own through iterative relational analysis.

This combination of quantum superpositions and kernels provides a massive speed-up by allowing the system to learn in fewer steps. Eleni Diamanti, a quantum communication expert with the Sorbonne University states "This allows you to show that you don't have to wait for the full-scale quantum computer, you can get the advantage out of quantum resources. You can already show it for some tasks today."

Delving into the abstracts of two papers published in Nature, "Power of data in quantum machine learning" and "A rigorous and robust quantum speed-up in supervised machine learning" can give some additional evidence that quantum computing is providing a speed-up, if not further accuracy, in machine learning algorithms.

Starting with __Power of data in quantum machine learning__: "we then propose a projected quantum model that provides a simple and rigorous quantum speed-up for a learning problem the fault-tolerant regime". Fault-tolerant regime meaning "the ability that allows a system (computer, network, cloud cluster, etc.) to continue operating normally without interruption even if one or more components fail."

Moving to the paper __A rigorous and robust quantum speed-up in supervised machine learning__: "Here we construct a classification problem with which we can rigorous show that heuristic quantum kernel methods can provide an end-to-end quantum speed-up with only classical access to data. To prove the quantum speed-up, we construct a family of datasets and show that no classical learner can classify the data inverse-polynomially better than random guessing."

It's clear that there is a wide range of researchers looking into quantum computing for machine learning and artificial intelligence applications. This proves fruitful for the zeitgeist of today with artificial intelligence being the front and center of today's technological innovations and popular topic amongst the mass population.

That's where Professor Andrezej Wichert is stepping in. He is showing people that you can use Qiskit to step into the realm of the quantum and apply quantum computing to machine learning algorithms, and artificial intelligence problems right now in the comfort of your home. Looking at his GitHub resources that will accompany his book, I can clearly tell that most of this stuff is not going to be for the everyday busy body. This is something that will take effort and modest dedication to truly understand. How his book will help give an explanation to these solutions is not yet seen but is highly anticipated when it releases in January 2024. You can __head over to his GitHub repository__ to take a look at what he has come up with and try to run these on your computer or on IBM's online Qiskit platform.

Overall, with the great success that Chat-GPT 4 has made on the minds of the populace and competition rushing into the fold, we can rest assured that faster ways of processing these models are going to come into the mainstream. Researchers have been hard at work to ensure their ideas become manifest and credit is given where it is due. Smart people are going to make these systems more efficient and user-friendly over time. The book mentioned throughout this post is hopefully going to give users some hope to enter into this fascinating world of the quantum and apply it to their favorite flavor of technology at this time, artificial intelligence.