DEV Community: Nina Chan

Here’s why Julia is THE language in scientific programming

Nina Chan — Wed, 08 Jan 2025 20:59:18 +0000

One of the best things about working in quantum computing is being exposed to a whole new set of skills and tools. For those of us coming from the software engineering side, discovering the things our academic peers are using is an interesting experience. One of the best for me personally was coming across the Julia programming language, and learning why it has become the go-to language for serious scientific programming in some circles.

What is the Julia programming language?

Julia is a relatively new programming language that has been gaining attention in the scientific computing communities. It was designed to combine the ease of use of Python with the performance of C. It also aims to be the go-to language for computationally intensive applications in fields like data science, scientific computing, machine learning, and over here in my world of quantum computing. Thankfully it’s not super difficult to get your head around and even easier to get started.

Julia was developed at MIT starting in 2009 and launched publicly in 2012 as an open-source project. It’s a dynamically-typed programming language built from the ground up for speed. Some of the key characteristics of Julia include:

Just-in-time (JIT) compilation using the LLVM compiler framework, enabling performance on par with C/C++
Optional type annotations, allowing it to be both dynamically and statically typed
Metaprogramming capabilities for code generation and optimization
Built-in package manager with an extensive ecosystem of libraries Familiar syntax, similar to Python and MATLAB Interoperability with C, Fortran, Python, R and more

Anyone talking about Julia usually points out how it is trying to solve the “two language problem”. This is where programmers use a high-level language like Python or R for prototyping and development, but have to rewrite performance-critical code in a lower-level language like C or Fortran. Which is something I feel in quantum computing where a lot of the machine level stuff is in C/C++. With Julia, the idea is that we can write high-level, expressive, and readable code without sacrificing performance. That’s why we see it popping up everywhere like:

Data science and machine learning: Libraries like DataFrames.jl, MLJ.jl, and Flux.jl are building a powerful ecosystem for data work. The ability to write NumPy and pandas-like code with better performance is a big draw.
Scientific computing: Packages like DifferentialEquations.jl and JuMP.jl are popular for modeling and simulation. Julia’s speed is a boon for compute-heavy applications.
Parallel and distributed computing: Julia has strong support for parallel computing, both multi-threading and multi-process.
Distributed computing frameworks like Dagger.jl make it easy to scale.

Getting started with Julia

If you’re interested in trying Julia, the best place to start is, funnily enough, the official Julia website at julialang.org. There you can download the latest stable release for your platform and find extensive documentation and learning resources. Some great beginner resources include:

The Julia Manual: Walks through the basics of the language with plenty of examples
JuliaAcademy: Free interactive Julia courses from Julia Computing
ThinkJulia: Free O’Reilly book teaching the basics of programming in Julia

https://www.youtube.com/watch?v=4igzy3bGVkQ

Once you have Julia installed, you can launch the REPL (interactive command line) or use an IDE like Juno or VS Code with the Julia extension. Create a new file with a .jl extension, and you’re ready to start coding.

Here’s the obligatory “Hello, World!” in Julia:

println("Hello, World!")

And here’s a quick example of Julia’s strengths — calculating the first 10 million prime numbers:

julia
function is_prime(n)
 for i in 2:sqrt(n)
 if n % i == 0
 return false
 end
 end
 return true
end
primes = filter(is_prime, 1:10⁷)
println(length(primes))

This code runs in just a few seconds on a typical laptop, thanks to Julia’s ability to compile is_prime into efficient native code.

So thats Julia!

And there you go. A quick taste of the world of Julia. Hopefully that gets you on the path to checking out more, or just helps show you why we’re learning yet another programming language in a world absolutely chock full of them. Given how familiar it feels if you’re already using Python and MATLAB, this hopefully isn’t too much drama, and puts you in a good position to be able to make the most of the high-level scientific expression and low-level performance that makes it such a kick-ass tool for technical programming work. Good luck!

These are the types of AI you need to know in 2024

Nina Chan — Wed, 14 Feb 2024 08:28:20 +0000

As children we thought of AI only as far as robots and science fiction. But now it's constantly in the news and being talked about as if it is one big thing. But it's not. There's a lot of different types of AI and it's important for anyone (let alone anyone in a tech job) to at least know what those types are and why the are different. So let's break down the ten main AI types you should definitely know about in 2024.

1. Machine Learning (ML): The Pattern Detective

When we talk about "Artificial Intelligence" lately we're mostly talking about some form of Machine Learning. As IBM say "machine learning is a branch of artificial intelligence and computer science which focuses on the use of data and algorithms to imitate the way that humans learn, gradually improving its accuracy."
The big point here is that machine learning systems get better over time. And this is powerful because we don't have to keep reprogramming them. I'm sure you've seen examples of Machine Learning systems being fed a ton of data (like pictures of cats and dogs). It figures out the patterns that mean "cat" vs. "dog" and gets super good at spotting them in new pictures.

2. Deep Learning (DL): Think Brain-Inspired Computing

Deep Learning is like Machine Lerning on steroids. It is a subset of machine learning that uses artificial neural networks with multiple layers to learn complex patterns and representations directly from data. If that sounds a little complicated, just think of these networks as something that is inspired by the structure and function of the human brain.
The key think here is that there are layers that the data goes through. Some layers handle certain classifications, some make more abstract and higher-level extractions, and the final layers are those that make the kind of output we have designed. This is how machines master games like Go, recognize your face (even for things like dating!), and understand what you're saying.

3. Natural Language Processing (NLP): Your AI Chat Buddy

Imagine teaching a computer to talk like a human? This sounded crazy even a decade ago (sorry Siri), but this is where we are at right now. Natural Language Processing is what powers the Large Language Models (LLMs) you keep hearing about in the news. This methodology helps machines make assumptions about what we write and say - powering those friendly chatbots, language translation apps, and even tools that figure out if customers are happy or grumpy from their reviews. Rather than just raw data, they are designed to analyse what we squishy humans say and write.

4. Generative Adversarial Networks (GANs): The Imagination Battle

You might not have heard too much about GANs recently, but they were all the rage a few years ago before LLMs took the spotlight. Think of GANs like an AI art duel! In this case, we can imagine two brain-like systems competing. One gets REALLY good at making stuff (fake images, music, you name it), the other gets REALLY good at telling real from fake.
The idea of making these networks battle each other (that's the adversarial bit) is that this effort helps create a continually improving system. While LLMs are all the rage right now, expect to hear a lot more about GANs for a while yet.

5. Computer Vision: Robot eyes in the Sky

Computer vision gives machines the power to see! Or at least that's the hope. The most noise about computer vision is coming from a certain car company that has been promising self-driving cars for a very, very, very long time. But there's also robots in factories and even those adorable ninja robots that are absolutely going to be haunting my nightmares tonight.
While it's probably not the smartest idea to build your entire car company around "computer vision is the only thing we need to work out roads and obstacles" for self driving cars, computer vision is essential in other systems like facial recognition and even medical imaging.

6. Expert Systems: Like a Rulebook for Problem-Solving

Let's get into some more focused use cases now. Expert Systems are a branch of artificial intelligence that focus on emulating the decision-making abilities of human experts within a specific domain. We can think of them as working off a big knowledge base and sets of rules.
The other side of this method is to use an "inference engine" to apply this knowledge. The inference engine is a component that processes the rules and facts against a user's query to provide conclusions or make some recommendations. Need to diagnose a weird medical issue? Ask an expert system for clues. They're great at recommending and planning too.

7. Robotics: AI + Moving Parts = Awesome

Here's where we get closer to either the T-800 or Johnny Five. Whatever our future fate looks like at the hands of our robot overlords, those robots get WAY smarter with AI inside. Think of Computer Vision for navigating, Machine Learning for learning cool new moves, Expert Systems for determining rules (like whether to hug or crush a human) and all of it working together - that's how they clean our houses and maybe someday will build stuff alongside us! Or destroy us! Yay!

8. Rule-Based Systems: Oldie but Goodie

Okay, they may seem simpler than fancy Machine Learning, but these rule-based babies are still around! They're those straightforward "if this, then that" decision-makers - useful when every step needs to be super clear. That also means that they are cheaper, and ideally are designed for very specific scenarios where variation isn't likely. Think of things like sorting machines, or… hey stop yawning. They can be exciting despite all the attention going to the killer robots.

9. Fuzzy Logic: Dealing with the 'Gray Areas'

Life ain't always "yes or no" or "black or white". Sometimes we need to be able to consider the various shades of gret. This is where Fuzzy Logic comes in. Think of weather forecasts saying it's going to be "kind of warm", or those smart washing machines figuring out just the right cycle for your messy clothes.

This is an interesting part of Artificial Intelligence because of the way the process works. The first step is to "fuzzifying" the inputs. This means taking crisp values like a temperature reading and mapping them to fuzzy sets representing linguistic terms. So things like "cold," "warm," and "hot." These sets of definitions have overlapping rules, but there are still sets of rules to apply to them, and an engine works out what to do. The fun part is that there's a process that "defuzzifies" back into a crisp value, like a specific fan speed setting.

10. Hybrid AI: The Ultimate Team-Up

The coolest stuff happens when different AI systems combine forces. While there's so much hype about LLMs and the various dramaas happening at LLM companies, the reality is that all forms of AI are evolving and crossing over in continuously changing ways. It's hard to keep up with but it helps to understand the basic types defined here to be able to look past whatever headline of the day, and think about what's really involved.
For example, a medical diagnosis tool might use expert systems for basic stuff, NLP to understand a patient's notes, AND machine learning to spot weird patterns in test results.

Quantum Computing for normal people: Grover’s Algorithm and the quantum search engine

Nina Chan — Thu, 07 Sep 2023 07:32:27 +0000

Here’s another post in my series on understanding quantum computing as a normal person. We might not be super normal to have an interest in this, but I started down this path as a way to ensure my career in tech is broad and long lasting, and I quickly realised that you don’t need a PhD to work in quantum computing anymore than you need an aeronautical engineering degree to work in cloud computing. Pardon the pun. But the point is, there are fundamentals we can be aware of as enough information to begin to contribute to an industry that needs as many talents and skills as possible. And that includes you too, so let’s jump in.

Introduction

Imagine going to the library to find that one book you desperately need for your research. If it’s a large library, you could spend hours searching through countless aisles and shelves. Now, what if there was a magical way to find your book faster, reducing your search time dramatically? Welcome to the realm of quantum computing, where something similar is actually possible! Specifically, we’re diving into Grover’s Algorithm — a game-changing way to search through databases quickly.

What is Grover’s Algorithm?

Grover’s Algorithm is like a turbocharged search engine but for quantum computers. Invented by Lov Grover in 1996, it’s designed to find a specific item in an unsorted list much faster than any classical computer algorithm can. While a classical computer would need to look through each item one by one, Grover’s Algorithm can find the answer with fewer steps, making it exponentially faster.

How Does It Work?

In a conventional computer, searching through a list is a simple but time-consuming process. Let’s say you have a list of 100 items. In the worst-case scenario, you’d have to check all 100 items to find the one you’re looking for. In the best case, you’d find it on your first try. On average, you’d check about 50 items.

Grover’s Algorithm, on the other hand, takes advantage of the weird and wonderful rules of quantum physics. In the quantum realm, information can exist in multiple states at once — this is called superposition. So instead of checking each item individually, a quantum computer could, theoretically, check multiple items at the same time.

The magic happens when Grover’s Algorithm amplifies the probability of finding the correct item while reducing the probability of the wrong ones. It’s like turning up the volume on your favorite song while muting all the others. After a few of these “amplification cycles,” the algorithm has a high likelihood of finding the item you’re looking for.

Real-World Applications

You might be wondering, “That’s cool, but why should I care?” Well, Grover’s Algorithm has some promising real-world applications. For instance:

Database Searching: Businesses with huge databases could find the information they need more quickly and efficiently. It’s a powerful application but perhaps less funded than…
Drug Discovery: Pharmaceutical companies could use it to sift through enormous chemical databases to find potential new drugs. A lot of quantum companies are talking about this now (as they seem to need to prove some kind of revenue potential).
Cryptanalysis: While this might not be the most uplifting application, Grover’s Algorithm could break some forms of encryption faster than classical methods. There’s going to be a lot of hype and noise around this. So take note and be sceptical until we see otherwise.

The Caveats

While Grover’s Algorithm is genuinely revolutionary, it’s essential to know that quantum computers capable of running it efficiently don’t yet exist. The hardware is still in the experimental stage, the funding seems to be drying up a little at the time I am writing this at the end of 2023, and scientists are tackling significant technical challenges despite the hype.

Moreover, the algorithm doesn’t make searching “instant.” It just makes it faster. For a list of 100 items, a classical computer might take 50 steps on average, but Grover’s Algorithm would take around 10. It’s a square root kind of relationship, so the bigger the list, the more time you save.

Wrapping up

Grover’s Algorithm isn’t just a fanciful concept from the realm of quantum physics — it’s a pathway to solving real-world problems faster than we ever thought possible. While we’re still in the early stages of quantum computing, understanding methods like this helps us peek into a future where we can solve complex problems more efficiently than ever before. So the next time you’re lost in a library, just remember: Grover’s Algorithm could be the magical solution we’ve all been waiting for. Now just to make these machines that can do just that…

Mastering Technical Topics: Effective Strategies for Learning

Nina Chan — Sat, 24 Jun 2023 21:44:38 +0000

Do you get this feeling? I've been a professional software engineer for a year or so now but I still feel like a student. And I always will. The pace of technology is moving so fast, whether that's Javascript or AI or my current studies in quantum computing. It's all moving so fast and we need to continually relearn how to learn. This is something I've been exploring lately.

Learning technical topics isn't always easy. So I'm exploring a more thoughtful way of understanding the right strategies to help make this consistent and regular. In this article, we will explore proven techniques that can help us learn technical subjects more effectively. Whether we're a beginner or an experienced learner, these strategies will enhance our comprehension and retention of complex concepts.

1. Define Clear Learning Goals

Before diving into a technical subject, it's crucial to set clear learning goals. I find this really hard sometimes as I want to learn it all and all at once. But that's not realistic. Determining what specific knowledge or skills we want to gain from studying the topic is essential. This will help us stay focused and motivated throughout the learning process.

2. Break Down Complex Topics

Complex technical topics can appear overwhelming at first. To tackle them effectively, we can break them down into smaller, manageable subtopics. Start with the fundamentals and gradually building our understanding by progressing to more advanced concepts is the way. This incremental approach allows for better comprehension and prevents information overload (although your mileage may vary!).

3. Engage in Active Learning

Passive learning, such as reading or listening to lectures, is often less effective than active learning methods. Actively engage with the material by taking notes, summarizing key points, and discussing concepts with others. I love having an information study group for this (especially on harder topics like quantum mechanics or how quantum computer work). Hands-on activities, like coding exercises or practical experiments, can deepen our understanding and help us apply theoretical knowledge. Or just be frustrating. Sigh.

4. Utilize Multiple Learning Resources

One is never enough. Diversifying our learning resources to gain different perspectives and explanations is such an important way to get multiple tones and contexts and learn widely over technical ideas. Apart from textbooks or online courses, we can explore tutorials, video lectures, forums, and interactive platforms. So many different places and ways to learn. Each resource offers unique insights, and using multiple sources can reinforce our understanding of the subject matter.

5. Practice with Real-World Examples

Technical topics are nearly always best understood when applied to real-world scenarios. I like to chase down any kind of practical examples that relate to the subject that I am learning (which is sometimes easier said than done). Working on projects or solving relevant problems is always the best way to enhance our ability to connect theory with practical application. It also promotes critical thinking and problem-solving skills. And just feels more fun at the end of the day.

6. Embrace a Growth Mindset

Adopting a growth mindset, and believing that intelligence and abilities can be developed through dedication and hard work, is so powerful. Don't get discouraged by initial difficulties or setbacks. Understand that learning is a continuous process, and mistakes are opportunities for improvement. Cultivate a positive attitude towards challenges and embrace the learning journey. All of this is easier said than done, but it builds up over time. I wish I focused on this earlier, it's such a secret super power!

7. Teach What You've Learned

Teaching others is a powerful way to reinforce our own understanding of a technical topic. Sharing our knowledge with peers, colleagues, or even online communities (like this one!) is a magical way to get that extra incentive to learn it all correctly, and so motivating to know it helps others. Explaining concepts in simple terms not only helps everyone else but also solidifies my own grasp of the subject (and I'm sure this is the case for you too). It's like magic.

8. Review and Reinforce

Regularly reviewing previously learned material reinforces our knowledge. I'm a fan of any system that can help with spaced repetition, where we revisit topics at increasing intervals, which is shown to significantly improve long-term retention. Using flashcards, self-quizzes, or online tools specifically designed for spaced repetition to make the review process more efficient is an overlooked learning hack. And one worth doing.

And what next? More learning!

Learning technical topics effectively requires a combination of focused effort, active engagement, and the right strategies. It's easier than that might sound. By setting clear goals, breaking down complex subjects, utilizing various resources, and engaging in active learning, we can enhance our understanding and retention of technical knowledge. A big part of this is remembering to embrace a growth mindset (even when it feels so hard to), and to practice with real-world examples, teach others, and review what we have already learned regularly. It helps so much to show how far we've come and a reminder that it's possible. We're doing it! With these techniques, we'll be well on our way to mastering any technical subject we set our mind to. In my case, that's quantum-based computing, but for you, that might be learning a language, learning to play an instrument, or just learning how to learn (so meta!).

Ten reasons why Quantum Computing is coming for Business (and soon)

Nina Chan — Sun, 11 Jun 2023 19:50:17 +0000

It’s never easy to predict the future. But the technology and business industries both agree (for once) that quantum computing is coming soon. Quite what “soon” means is probably less agreed on, but the areas that they know will be a first wave of solutions and products coming through is pretty clear. And worth us taking notice of to make the most of this new technology and the massive global change that will follow.

1. Unmatched Computational Power

Quantum computing leverages principles of quantum mechanics to process information, enabling exponential computational power compared to classical computers. This is particularly useful for complex problem-solving and it is likely that quantum processors (as QPUs) are an addition rather than a replacement to GPUs.

2. Advances in Cryptography

Quantum computers can revolutionize data security. There’s a lot of noise around quantum computing breaking encryption, but the other side holds true. Being that any breaking of one form of encryption usually creates another era of it. And more solutions to come. Quantum cryptography and quantum key distribution could make data transmission virtually un-hackable, a crucial advantage for businesses.

3. Efficient Data Analysis

Quantum algorithms can analyze and draw meaningful insights from massive datasets far more efficiently than classical computers. This will significantly speed up data-driven decision making in businesses. This will especially affect the industries heavily reliant on complex or time-sensitive calculations, such as medical research, supply chains, or even the tax department. Oh no!

4. Improved Logistics

Quantum computing can optimize complex logistics problems, like route optimization and supply chain management, thus reducing costs and improving efficiency in businesses. This will likely be one of the big initial areas, and the routing of vehicles is something that has always been present in computing breakthroughs right back to the first digital computers!

5. Machine Learning and AI

Quantum computing can accelerate machine learning algorithms and artificial intelligence applications, enabling faster, more accurate predictions. this will be an area where there might be a lot of fuss and noise and hype (especially given the current industry fixation), and no doubt there will be a lot of over-the-top claims. But the field is still one with real overlaps with quantum computing.

6. Drug Discovery and Healthcare

Quantum computers can simulate molecular interactions at an unprecedented scale, promising breakthroughs in drug discovery and healthcare — a lucrative avenue for businesses in these sectors. The market size here is huge, and any small advanced have a massive upside, so we can expect a lot of industrial application in this area.

7. Material Science Advances

Quantum simulations can lead to the discovery of new materials with desirable properties, opening up opportunities for businesses in various sectors like manufacturing, energy, and electronics. And given the state of the world (and the world economy) maybe not a moment too soon.

8. Financial Modeling

Quantum algorithms can optimize portfolio management, streamline financial forecasting, and enhance risk management, providing a competitive edge for businesses in the financial sector. Fintech bros rejoice.

9. Climate Modeling

Quantum computers can process complex environmental data to model climate change impacts, enabling businesses to plan for sustainable practices and comply with environmental regulations. The results might be surprising, given we might not like what we get back, but using tools like this for advances or insights into saving our world is a valid use surely.

10. Growing Quantum Ecosystem

More tech giants and startups are investing in quantum computing research and development, indicating the imminent arrival of quantum computing in mainstream business operations.

These advantages are real and coming. The timing is harder to predict as the industry seems to swing from confident to cautious. But it’s enough to say that the world of business had better be paying attention. Or be left behind.

How to thrive (and survive) your first week as a new software engineer

Nina Chan — Sun, 11 Jun 2023 08:32:24 +0000

Congratulations, you've landed a new job as a software engineer! As you get ready to dive into your new role, the first week can be a whirlwind of information and new experiences. It's common to feel a mix of excitement, anticipation, and a bit of apprehension. To help you navigate these initial days, here are some valuable tips to make your first week on the job as a software engineer a resounding success.

Get to Know Your Team

Your team is going to be your main source of support and knowledge. Take the time in your first week to get to know them. Understand their roles, how your role fits into the team, and how you'll be working together. Don't be afraid to ask questions, and show genuine interest in your colleagues' work and experiences. Building strong relationships will set you up for a collaborative and productive work environment.

Understand the Codebase and Tools

As a software engineer, you'll be spending a lot of time working with code. Your first week is the perfect time to familiarize yourself with the codebase. Start by reading the documentation, checking out the code, and understanding the structure and the patterns used. Don’t hesitate to ask your colleagues for pointers if you're unsure where to start.

Also, get to know the tools and technologies that the company uses. Whether it's a specific integrated development environment (IDE), a version control system, or a particular framework, take the time to understand how these tools work and how to use them effectively.

Learn the Processes

Every organization has its own set of processes and methodologies, whether it's Agile, Scrum, or something else. Understand the workflows, how tasks are assigned, how the team handles bug tracking, and how they conduct code reviews. Knowing these processes will help you fit into the team's workflow and contribute effectively.

Don’t Be Afraid to Ask Questions

Starting a new job can be overwhelming, with a lot of information to absorb. If you're unsure about something, don't hesitate to ask questions. It's better to ask for clarification early on than to make assumptions that could lead to mistakes down the line. Remember, everyone expects you to be learning in your first week, so they'll be ready to help.

Take Initiative

While it's important to listen and learn, don't be afraid to take initiative where you can. Whether it's offering to help on a small task, suggesting a new idea, or asking if you can shadow a more experienced engineer, showing initiative demonstrates your enthusiasm and can make a positive impression.

Take Care of Yourself

Finally, remember to take care of yourself. Starting a new job can be stressful, and it's easy to neglect your health. Make sure you're getting enough sleep, eating well, and taking breaks when you need them. Balancing work with self-care will help you stay focused and productive.

Conclusion

Starting a new job as a software engineer is an exciting time, filled with opportunities to learn and grow. Your first week may seem daunting, but remember that it's a time for learning, asking questions, and getting to know your new environment. With a positive mindset, a willingness to learn, and these tips in mind, you're well on your way to a successful start in your new role.

My quantum leap: blogging while learning

Nina Chan — Thu, 08 Jun 2023 01:29:20 +0000

Hey there! So I've got some news to share. Hitting my mid-twenties I decided that it's time to take the plunge into programming, and not just any programming - but quantum computing. This is kind of crazy but with a bit of physics background I decided it's a good a time as any to start learning about the world of quantum! As I dive headfirst into all things quantum, I've decided to write about it right here on this blog.

Think of this blog as my personal diary, where I scribble down all my quantum adventures. It's a reminder of how far I've come and how far I've yet to go. Blogging while learning has some serious perks:

Making Stuff Stick: When I write about something, I've got to really get it. It's one thing to read about spooky quantum stuff, and another to put it in my own words!
Thinking Cap On: Blogging is my thinking cap. It helps me ponder, figure out what's working and what's not, and tweak my learning game.
Chit-chat: My blog's a great place to meet and chat with other quantum buffs. I'm hoping to learn from their stories and maybe even make some friends along the way.
Show and Tell: Over time, this blog could be my show and tell, a way to show off my quantum chops. It's proof that I can hang tough in the quantum world.
Sharing is Caring: I'm a big fan of the saying, "When you teach something, you learn it twice." By sharing my quantum escapades, I hope to help other quantum newbies feel a little less lost.
Standing Out: Blogging's a neat way to stand out in a crowd. It's my chance to put my own spin on things and make my mark.

So, there you have it. This blog is my promise to myself to dive deep into quantum computing and come out the other side. I've got my keyboard, I've got my curiosity, and I'm all in.

Care to join me? Let's make quantum computing less of a mystery together. Bring on the qubits!