Transcript
Rene: Hi welcome to QuBites, your bite-sized pieces of quantum computing. My name is Rene from Valorem Reply and today we're going to talk about data handling for quantum computing. And for this, I am honored to have our expert guest today Spencer Cook. Hi Spencer, how are you today? Welcome to the show!
Spencer: Hey Rene! Thanks for having me on, I've always wanted to be a guest.
Rene: Awesome, love it! Well, you can already actually provide real value, but [first] tell us a little bit about yourself and your background as relates to data handling, quantum computing, whatnot. And also, I know you actually have your own YouTube video series, right?
Spencer: Yes! Yes, thanks for calling that out. Yea, so, I'm a Senior Data Scientist at Valorem Reply. Rene and I are coworkers. I'm on the Data & AI team. My first interest in quantum was actually spawned by a professor I had in undergrad named Jeffrey Uhlmann at Mizzou. He just happened to be a really, really proficient guy in quantum computing. And like you mentioned, I do have a weekly YouTube series where I cover Databricks, quantum computing, all sorts of new things in the Azure data space. So you can check that out on the Valorem Reply YouTube channel.
Rene: Love it, love it! Well, let's dive right into today's topic! Since you’re the data expert, right, let's talk about what are the challenges when we're dealing with data and quantum computers? Like in terms of maybe storage, but also you know memory, and all of that. I know there's a bunch of challenges, so what are the challenges when dealing with data on quantum computing?
Spencer: Yeah. So you know, Renee like you said, it's hard to pick just one. But the biggest challenge that I envision whenever it comes to customers I work with and their quantum future, is how data is stored and managed right now. So, there is a lot of basically bad data hygiene that's become kind of the default because of things like cold storage, because of you know cheap, lazy evaluation and big compute environments like spark. And I think people are going to start to see a need to go ahead and get their data ready for quantum computing ahead of time so that there is not a bottleneck with hardware resources once we're actually you know ready to go, and you know, there's more availability of quantum hardware.
Rene: Gotcha and so one thing you’re saying is how you can address [data challenges for quantum] is like plan ahead. Its maybe similar when we think about quantum security and quantum computing threat in terms of cryptography, right? Where we have these encrypting algorithms, like RSA for example, it can be cracked in a few years once there’s enough qubits available and so on with prime factorization and all of that stuff, right? Which we, by the way, covered in a previous QuBites episode if you want to check it out. The point I'm trying to make is, basically, when we think about planning ahead with quantum security, like the threat with post quantum cryptography, planning ahead so that you can quickly for example exchange your certificates, a thing called crypto-agility and whatnot. And so, can you do something similar with data? Is there data agility, or what is the approach [for achieving agility with your data]?
Spencer: Well, so one [data preparation step] that immediately comes to mind is feature stores for machine learning. So, instead of storing your entire data set and going ahead and preparing that to be stored in the memory of your quantum computer, quantum ram; which requires a good deal of overhead right to do that conversion from a traditional memory mechanism to a quantum one; you can instead just translate your features. So, if you've already pruned your data, if you've already investigated [to determine] ‘this is what's relevant within my data set’, and then you just move that subset over, you're going to have much more agility and spend a lot less on your quantum resources.
Rene: Gotcha. Gotcha. What else can customers do to be prepared with their data for quantum computing?
Spencer: So, other than just getting in the habit of doing things like storage vaults or future vaults rather, I think using things like Azure Purview, is going to be huge as well. So, basically, just having a top-down view of what data you have available, what it explains - again you don't necessarily want to spend the time doing your investigating on your quantum hardware, you want to have this investigation and labeling done. So, all the normal things we say about data governance and good data management in general, become even more critical in the quantum space.
Rene: Right. So, basically, take care of your data and prepare yourself with your data, not just for the quantum future, but also for getting more insights out of the data. And of course [Valorem Reply’s data experts] can also help along the journey, right? Well, we’re already at the end here. So, thank you so much spencer for joining us today and sharing your insights, it was very much appreciated.
Spencer: Yeah, absolutely! Again, thanks for having me on, I'm always excited to talk about these cutting-edge things like quantum computing.
Rene: Awesome! And thanks everyone for joining us for another episode of Qubites, your bite-size pieces of quantum computing. You know it, I will say it, watch our blog, follow our social media channels to hear all about the next episodes. And also, of course, subscribe to the Valorem Reply YouTube channel, where you can find Spencer’s series. He's doing live sessions every few days or at least every week!
Spencer: Yep, and feel free to also join the Valorem Databricks User Group (VDUG) [on our YouTube Channel the last Tuesday of each month].
Rene: There we go, right, databricks! Alright, so again, thank you so much everyone for tuning in. Take care, be safe and see you soon! Bye, bye.
