Can we Eat a Computer??!!
To eat, we must think first. Part one of quantum food series
So if you clicked on this article wondering how it’s possible to chomp down your MacBook during lunch, then I’m sorry to tell you, but [⚠️DON’T DO THAT!!] I won’t be covering that today. However, there have been some mukbang-ers who I’ve seen eat iPhones… Anyway, edible chocolate devices isn’t what I’m going to be talking about today.
So if you’re not any type of food -tarian, and you’re not a super particular eater, then you probably saw those foods in the photos and recognized them as some of your favors. Well, this isn’t a bad thing. Dairy and meat are both just so good! But another thing they’re good at aside from taste is accelerating global warming. No, it’s not really the meat itself, but rather the way in which we produce it.
The production of dairy and meat from animals possibly helped to usher in the extinction of over 33 species, cause cancers or high cholesterol levels, and make up 2.9% of all anthropogenic greenhouse gases. In fact, by just cutting down meat production to nutritional needs only, we would’ve been able to save land 1.5 times the size of the European Union, and prevent 19% of water pollution from dairy. Now, it’s too late to just stop or decrease production, or get people to become vegan. But what if I told that by using quantum computational machine learning, we would be able to still eat meat, and completely mitigate these effects?
So if you want to better understand what I’m about to tell you check out my other article here, where I talked about how quantum chemistry and protein analysis can be completed using the VQE algorithm. We can still do the same, which I will tell you in the next part of this series, but we first need to jump over a huge issue. Literally.
Since I’ve started newer projects in the intersection between AI and food technology, I’ve been wondering about how I can incorporate another one of my 4 core interests (AI, Quantum Computing, Biocomputing, Cellular Agriculture) into this project for a higher quality and optimized product. Then I thought about it. We’ll I’m already using quantum computing for chemistry and security, so why can’t I repurpose that for chemical proteomics, or the study of the chemistry of proteins?
Well, I can. Just like how we can use something called quantum annealing to help in solving the protein folding problem for acceleration of the drug discovery pipeline, but we can do this in food too.
The Protein Folding Problem
Foods are comprised of multifarious chemical substances that can be categorized into two models:
- Macronutrients: Quite large quantities that make up a large amount of the nutritional value in food. Some of the macronutrients in meat and dairy include casein (dairy), and myofibrillar proteins (50–55%, mostly myosin and actin), sarcoplasmic proteins (30–34%, mostly enzymes and myoglobin) and connective tissue (10–15%, mostly collagen and elastin fibres embedded in mucopolysaccharides) for meat (all of macronutrients are mainly just important proteins).
- Micronutrients: Main nutritional substances that are found in small quantities, just as water, vitamins A, D, B12, B2, and B. Many B-complex vitamins are present in meat and dairy.
Vitamins are essentials structures made by proteins. With all of these proteins stored inside of food, we encounter the protein folding problem when considering edible materials on a classical computer.
Protein folding is the physical means by which a protein acquires it’s three dimensional structure. Essentially different reactions to water repellence — hydrophobic — and water attraction — hydrophilic — help to induce the folding process which helps in developing the final protein structure, as shown below:
A full protein structure is extremely important in determining different protein functions and ability, such as interactions, receptors, additional binding, and channel quantity, all of which make proteins so vital to maintaning life. If this folding process is corrupted, it can create prions, or misfolded proteins that are pathogenic and can spread their misfolded shape to other proteins. This gives rise to a host of neurodegenerative prion induced diseases like Creutzfeldt-Jakob’s disease (I linked the narrative I wrote on that) and Kuru.
The intricate process of protein folding is where this quasi-computational problem comes from.
The protein folding problem essentially describes this question:
How do amino acid sequences (the order of the building blocks of proteins) correlate to the expression of its 3D structure?
When I was first exposed to this problem, I was wondering why AI or a normal computer couldn’t solve it. After all, it was introduced in 1960, and still somewhat remains unsolved. This is due to the many causal factors and different reactans. In addition, conversions happen concurrently, and proteins have weak interactions. Clearly, this problem will not be an easy one to solve, and even things like deep reinforcement learning (DNL) and other AI algorithms have been partially unsuccessful, and obviously our classical computers aren’t do to well.
So how do we solve it so that we can synthsize new foods. Well, we take a quantum approach.
Solving 👆🏾w/ Quantum Tech
So clearly there’s a problem, but it’s already been identified, so step 1 is complete. Instead of the scientific method— sorry to every science teacher or Aristotle (but he also used first principles) seeing this… — we’re going to use this unique first principles approach that I’ve constructed:
- Have quantifiable and qualified assumptions
- Create or observe an actual problem that can be solved or tested
- Identify possible candidates for a solution
- Create, simulate, give or at least talk about a solution!
So right now, we’re somewhere between step 3 and 4. The way we can help solve the protein folding problem is using a meta-heuristic, or a powerful computational solving method, called quantum annealing, or QA.
Quantum Annealing as a Concept
So let’s make this quick.
In the simplest, but most intuitive and useful way, quantum annealing basically takes a problem with numerous random state changes based on a loss function. If you don’t get it, it’s fine. Let’s apply it to the protein folding. When given:
You get the absolute minimum of the objective function, despite there being multiple local minima, or the smallest value point within a given range of a graph. Quantum annealers are mainly applied to a mathematics field known as combinatorial optimization, where thee given space of a function is finitely discrete. Some problems that QA is geared to solve include the NP-hard problem known as the traveling salesman, a magneto-physics problem called the ground state of a spin glass, and of course the protein folding problem.
If there’s anything you take away from this section, visualize it in a physics conceptualization. Qubits operate in a state of superposition, in which they assume both binary (1,0) values:
The quantum annealer gets a complex graph function, and determines its absolute minimum utilizing quantum mechanics and superposition. I know that’s a little abstract, but it gets the job done.
Finally, let’s talk about why a normal computer can’t do this and why a quantum computer can and can’t.
First off, this is a quantum process. We’re considering thermal/quantum fluctuations, which a normal computer can’t do. Period. Sorry iMac Pro!😭🖥
Well, QA’s actually have their own classical counterpart, called simulated annealing, or QS. Simulated annealing is an optimization method, after explicit problem solvers like D-Wave, who takes in a problem using an energy statement,
and then the qi ∈ {−1,1} in the Ising, or qi ∈ {0,1} in the widely perceived quadratic unconstrained binary optimization (QUBO) model. After all of the necessary problem construction and recalculations are made, determining the given matrix problem is then passed to the simulated annealer for optimization.
A simulated annealer is able to do this by using the Equation (form described above), and is able to avoid being trapped in the graphical optima by using energy surfaces to move through the peaks, and may acquire additional enumeration steps to find the lowest energy solution.
A quantum annealer instead takes a wholly different approach, as previously described, and is able to find the lowest energy solution much faster, as it doesn’t need to climb through the whole function! It does this using quntum tunneling, another emergent qubit phenomena based on their weird state properties, and it basically phases through the energy wall, much like Shadowcat from X-Men (Marvel people, where are you). This allows them to move through the space between the local minima without any optima being disrupted.
Quantum Annealing as the Solution
So apparently, all we would have to do is pass the protein folding problem to the quantum annealer. Yep! It would hypothetically work. Currently, most of us will have to stick with simulated annealing and other quantum methods as we continue, but I’m currently developing solutions to expedite cellular agriculture using quantum computing!
After we solve the protein folding problem, we would be able to determine new protein structures, and what amino acid sequences we could use to consttitute them, along with a cohort of even more advanced genetic algorithms for building such proteins. Not only will this revolutionize cellular agriculture proteomics for milk and dairy, but quantum annealing as a lot of other practical uses, especially in the dynamics of everyday life!
For milk and dairy, this would mean healthy genetic engineering. In cellular agriculture, we give myoblasts or yeast the necessary injected DNA and proper environments to grow and thrive. But, we haven’t really seen quantum computing being employed to help in this process. In fact, this could completely change the architecture of the cellular agriculture field. We could engineer new proteins for food, increase nutrition, customize food content, change “food cell” compositions, and more! This would also make it easier for us to engineer lab food on a macro scale, and provide new knowledge in more disciplines, like genetics, additive manufacturing, public health and healthcare/nutrition.
We live in such an amazing world of new inventions and constantly expanding knowledge. We need to use that to solve some of the world’s biggest problems, and tackle the deepest and scariest issues that we see will serioiusly disrupt the way we live. Well, if we’re going be disrupted, then our systems will too. The only way to stop that is to attack and disrupt the problem back. So let’s make an impact on one of our biggest (and tastiest) problems. We can solve global warming, food impoverishment, famine, and world hunger. Let’s disrupt food as we know it.
Who knows, maybe in the next few decades, our children won’t even know what farming on a field actually is…
My name is Okezue Bell, and I’m a 14 y/o innovator/entrepreneur in the quantum computing and AI spaces. I’m also currently making developments in foodtech and cellular agriculture, as well as biocomputing! Contact me more:
✉️ Email: okezuebell@gmail.com
🔗 LinkedIn: https://www.linkedin.com/in/okezue-a-...
📑 Medium: https://medium.com/@okezuebell
🌍 TKS Acc: https://tks.life/profile/okezue.bell
📱 GitHub: https://github.com/BellAI-Code
💻 Personal Website: https://okezuebell.squarespace.com [password is 0; squarespace purchase not working lol]
