IPFS IPFS Camp 2022 - Decentralized Science (DeSci), 2 Nov 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Transforming big scientific data - Carla Ostmann

Description

This talk was given at IPFS Camp 2022 in Lisbon, Portugal.

A

um So today, I'd like to talk about what we're building at DSi Labs, which um yeah is going to be one of those d-site apps, that's building on all the cool infrastructure that everybody else is building here and particularly I want to talk about how our new program called node stewards is going to help transform big scientific data into intraoperable research objects that are stored, open State on this cool infrastructure that we just heard about so um yeah using those ecosystems really cool. It's popping lots of cool projects.

A

Many more I forgot a bunch, but yeah Jocelyn is. We are always curating. Those cool um landscape so just check it out. I have the Twitter right there and yeah, so we just heard about it. So it's sharing scientific data is super important. Why? Because? Well, if we share data, we can collaborate much more easily.

A

We can build bigger data sets and bigger data sets means more statistical power, more reliable results right, so that's pretty cool and it also means more access to the data that so there's not the same access to cool instruments that help you with data collection across Labs. So if you're in an underfunded research institution, you just may not have the ability to collect the same type of data that a well-funded institution may have. So if we all share data, we all have better access to make cool scientific discoveries.

A

So that's pretty cool right, but also sharing scientific data. Right now is pretty expensive. It's pretty vulnerable because it's stored on centralized databases where we just have to trust that they keep their database running. It's also not rewarded So, currently, really what counts and science is having your PDF cited, but it doesn't matter if you make your data accessible like you. Just cannot accrue credit to it. Are there some ways you can, but it's just not really easy, and it's also pretty painful. So there's a couple of repos out there. We can store your data.

A

These are like funded by some governmental institutions. There you access, not great and then also, if you want to find the data you need to know which repo it's stored at. So you need to find the repo. Then you need to find the data. It's all it's a hassle, so it's not great. So what I'm hoping to convince you of is that sharing scientific data is great if you're using these side nodes right. So what is design notes?

A

That's what we're building at DSi labs and basically it's a way for you to collect all your artifacts that are relevant to a specific piece of research that are currently scattered across the internet and different siled data repos, but also maybe not even be published, and you can collect all that in one DSi node make it you can enter. You can connect all the different artifacts that are relevant to This research. You can connect your code to your data.

A

You can link that all to the relevant sections in your PDF in your manuscript, and that way you can create a reproducible, research object and the cool thing is then, if you publish it, it'll become part of this open state repository that is based on the centralized Storage Solutions, where you can even have different nodes interact and one node, for example, citing the data set of another node, citing the code analysis of another node and then coming up with new results.

A

Based on that, so it's a very interactive scientific ecosystem, which that's kind of how it should be I, feel um so yeah I'd love to show you guys a demo. um So let's hope this works right. So the idea is that that's how you find knowledge right now right, so you go to some journal's website. You look for the PDF and then you click on it. You find a PDF. So this is how research currently gets shared.

A

You have this PDF, it's static, you're, not really you're, lacking a bit of context, so somebody could download this and send it to you to read it, but you can't really tell how has this been updated? Is this up to date? Has it maybe even been retracted and with this PDF, it's kind of cool, because you have this little button right here, which will always take you back to design nodes where you have all the contacts that's relevant to this piece of research, so you get the idea right. So you have interconnect.

A

uh That is running on cool projects like Buckley all um yeah I'll show you the a demo. If you come to me later, this kind of sucks anyways so now I want to talk about node, steers notes. Yours is our community program with which we're hoping to get actual scientists into the world of DSi show them what we can build if we use decentralized uh solutions for science and then hopefully convince them of the whole value proposition.

A

So note, stewards are are programmed to help scientists, onboard entities, I use this new technology and they're going to do that by creating the first 100 DSi nodes and so I'm going to walk you through how we're trying to implement this project. So this is step one, that's where we're at right now, so currently we're in the phase where we're trying to identify the most valuable research.

A

Was that mean that means we're trying to find research that will have the most immediate benefit from adopting these Solutions, and so that's likely going to be empirical work with large data sets with non-destruct restricted data, so we don't run into gdpr issues and hopefully a high impact research.

A

Just because it's better visibility, so once we have identified the most important field of research and the course researchers who are willing to maybe try this Tech we're going to collaborate with them, we're going to get their consent because, yes, we're looking for research, that's open access and permissively licensed, but we still want to get their consent because we want to work with them and not step on anybody's toes and then we're going to collaborate closely with them to teach and learn.

A

So we're going to teach them about our Tech and we're going to learn from them about their research areas, their problems, their issues and how we can help them and then, together with them, we're going to start collecting artifacts so collect the code. Collect the databases collect anything else. That's related to their research that we want to put up on DSi nodes- and this is important because currently, as I said, much of the artifacts are relevant to research are not necessarily publicly available and that's what we're trying to fix.

A

So we're going to work with Scientists to identify those artifacts and then once we're there. Next step is to actually go and create these 100 Stellar design nodes. So we're going to connect all the artifacts, that's going to enable one-click reproducibility, which you will know exactly what I mean if I was able to show you the demo and we're going to also create artwork that we're going to tie to This research just to make it a bit.

A

You know make it a bit more cool and then we're placing high value and making all the research that we publish on design, notes, fair and fair means, findable, accessible, interoperable and reusable, and that's going to lead to good machine readability which just helps you find the research that you're looking for find the data set that you're looking for and just use all the artifacts that are already out there to then better, create your own research.

A

Then we're going to publish all that through the ipfs network and then we're going to let the authors of the work claim their work as author us and yeah so Step Zero. This is a constant step that we're gonna be implementing. All the time is just building community. So we have a couple of really engaged community members which who we're super grateful for who are willing to help us with this project and create the first nodes and yeah, so we're onboarding stewards on a rolling basis.

A

So if you're interested hit me up and then we're facilitating connections between these stewards because there's so much opportunities for us to teach and learn from each other. So we have a really cool community of people who are web free native scientists. We have actually data stewards themselves whose actual job it is to work with researchers to understand and help them implement the data requirements that they have for sharing the data. We have artists, so we have a bunch of cool people and we can just learn together and have a good time together.

A

Well, we nerd out on Research, so it's kind of cool, um so yeah. What did we learn in the process? Actually, there's a bunch of people who want to fix science and I think that applies to many of you here in the room and we can leverage that momentum. That's there by just giving these people the tools and the means to then actually go ahead and help starting to fix science.

A

um Another point that we learned is that if you have a diverse Community from different fields that just helps you understand your users a lot better. So we have been talking to our community and just most of them are scientist. So we learn a lot about what are the specific problems that they're having with sharing their data? So that's super super helpful and yeah. Just in general, communities are super powerful and we need to nourish them. So let's do our best to do just that um so yeah and then back to the ecosystem.

A

um So if we onward scientists um with this note shares program that just helps the entire ecosystem because, as I said, there's so many cool uh projects out there. So if we get more scientists in the space, um then they're gonna use all the other cool projects and we can all help each other and work together and doing that so yeah. If you're interested in becoming a note, Stewart join us, uh you can go to DSi foundation.org stewards. You can also reach out to me or follow DSi Labs on Twitter. Thank you.