 Okay, yes, thanks for still being here and I will try to hurry because I also have a train to catch Yes, I'm a postdoc at the Hamburg University of Applied Science started in summer in June, July and We have a Growing group that is working with blockchain and we are about 15 people now and Mostly we have IOT projects mud city projects, but we also have a Project that I will present today and it's a bit or it start as a project from the research re-use of research data and we During the the past month. We thought we should extend this to open science, but Yeah, short introduction. I will present this research project could be T that Then I will show the architecture that we have come up with so far and our use cases that we have developed so far and are now in the way of implementing those so the research project is funded by the BNBF and We are three partners. We are the F EZ Karlsruhe, ENP Geiswald and we from Hamburg and All three groups at different parts to the project. So We have Karlsruhe. They are working on ontologies so that we can semantically interpret the research data and Also with a goal to to find this data and also to reuse it Interdisciplinary this interdisciplinary And EMP Geiswald, they are actually doing the plasma science. So the project is based on plasma research and They have the the problem that they do many experiments, but the results are not often reused not so well documented and But this is actually one area where it could be very You could benefit from reusing the data from other experiments, but since not all the metadata are there, for instance, what are the What is the machines used for the experiment? What is the environmental condition? So that's why Research data is often not reused, but people just do the same experiments again and again so this is actually the main motivation behind the project so that the experiments don't have to be done again and again, but we can reuse the data and Importance there also that we have quality criteria so that we can Identify which research data which results are on a search in quality level so that it really makes sense for me to reuse this data and With a blockchain technology it comes a bit together So we would like to use smart contracts that automatically do this quality check and We want to keep ownership However, as I told we thought okay this Concept that we came up so far with they would be interested even in a larger scope in the scope of open science And also of publishing so actually this talk would have fitted very well in the session with Blocksberg and Artifact there are probably some similarities or there are quite some similarities But so our main question was so there are many open access journals around or there are even Completely free journals around where you can publish for free you get reviews, but these journals are often Not used but rather people use the well-known journals pay money to have open access articles so And to the point is We thought so what are the steps that we have to take to to foster open access in a better way and also the open journals or some alternative structure to publish And said so the first thing so this applies both applies both to data and to publications so Yes, we need to secure ownership and data integrity So if we talk about reuse, it's also this point So when someone reuses your data, it should be somehow ensured that that person really reused your data without Changing it slightly to make it fit better for his or her own research then second is Data has to be findable accessible and the quality has to be ensured as I said already, but the main point is the third one Why are people publishing with? Expensive journals because they have a certain reputation and if you publish they are the thing okay Your research has a certain standard and when you get cited you have certain reputation. So It's similar to the artifact approach. You have already this incentive of citations and this would apply As well for data and if your data is reused it is more or less cited by your own research So we somehow need to ensure this incentive and then we believe if we get a good structure That we can give this incentive with our solution then people would Go rather with Alternative solutions for open access then with the common journals This is the architecture we came up with It's three parts the researcher who is author the researcher was reader and the database blockchain part in the middle So it starts with the researcher doing the experiments the data will be transferred to the research database locally then the researcher can use a web UI to Yeah, access our services from our blockchain solution and to first for instance Yes, certify the research data. So then we write the important metadata and the hash to the blockchain and other metadata will then be published by a conventional database So for as the excess to you search and find your research and when Other researchers will then why the web interface First search for data and when they find something interesting or for publications they can ask for access Get the data transferred either Yes, why a new link from the original research data management to their own Then it can be hashed again to ensure that really the right data has come over So we have these services force for Dapps. We have to certify verify Contract for one for publishing one for access curation, which is the quality control then review process so I Guess when you have public Publication for free search papers and the review is kind of the quality Control and the reputation management however We haven't really implemented this review case, but as we have seen L and I is working on these Fards so it fits very well together So I present three use cases This is the the simple use case that all of you know and that is nothing new just to yeah request a certification when you have When you're done with your research In the more complex sequence diagram I hope it's not too technical, but this is what we use then for the implementation follow this So it will be this step. So the researcher gains data with the Sensor application from the experiment and this data is stored in the research database so in the perfect world We could really Already access the data from the sensor and creating the hash directly here So that there's no way that this data is in any way modified But this Yeah, we are not there yet then in the next step the researcher would ask for a certification and Here is what's also Something that is a bit special of our solution But what also artifact is working on now and was also our first Challenge that we really sent the hash algorithm back to the research database so that the hash is actually calculated locally But from an algorithm that we control wire or not we but that the blockchain solution controls So that we really ensure yes, the data is there. Yes, the data has been hashed with this algorithm and we have also developed a kind of solution that We've talked about the case there was some question from the audience So what happens in five to ten years when actually the hash algorithm is not it might not be secure anymore Maybe we have quantum computers that do it completely differently So yeah, we also implement an algorithm that can update the hash algorithm later But we then come to the problem that we might deal with huge data sets for instance in physics they have Terabytes of data actually or at least gigabyte. So that's another challenge For the next use cases. I won't show these Sequence the earth because it comes very too detailed and you wouldn't be able to read it anymore. I stay on this More on a flow chart level So when we try when you want to have a publication So the first use case was just to certify the data But here you actually publish your data you want to make it public And this could be research data that you have already created a hash or it could be a publication That is based on other research data either your own or others or it could cite other documents other publications so your request your publication and The publish contract will try to verify the objects that you have used so it checks the Your citations so is there any data involved and then it can The certifier can ensure the hash is so you have to submit your Your research item and use data set so that you can ensure the hashes check in the blockchain if The use data set still has the same hash as entered in the first place And if this is done Then you can continue to the quality check other ways you will get a rejection and Then the quality check will do the curation. That's the methods developed by imp rice flood And if the quality is Okay, so you can come up with different types of quality So the the easiest thing would be is the data accessible is the data It doesn't have the same hash, but you already have done that is the meter the meter data complete And then you can also have some sematic analysis of the data So if you have the quality check done, it could also be the review process in case of publications Then you can publish the item You will save The the hash and the meter data the relevant meter data So if you think about these experiments, for instance, it could be important What temperature has been used when you bought these results? So we would create a hash of this kind of important meter data So when you want to reuse the data, you can be sure that this is ensured this kind of environment Yes, so you save it and then we go to the reputation management and increase the reputation so here we have the and the challenges that we really need to Identify the use data and we somehow have to extract if it was really I just noticed a line there, so I Shouldn't step over it. So Yes, so When we have a publication it will be hard to actually see which data set from the original dataset has to be used so This is an interesting challenge and we might find a solution for that one Or to really ensure that for this publication this dataset has to be used because there's this transfer That you do when you do research whatever So Yes, exactly and the third use case is the excess Where another researcher finds your work Ask the meta database to return the object. What would be a link to the research data and a link to the transaction in the blockchain so The reader will request access and this will be prepared So the blockchain will resolve the link and if it's more private data. So if it's still We think also of the so you have public data and you have more or less private data where you can control the excess So in the case it's private data You have to grant access first. Otherwise, it would do it directly and just ask the system the research database from the other from the author Create a copy from the data. This has to be hashed again locally and Yes, then we will check is this the same hash. So you really got the data Then it's a success. So we're done and we increase the reputation. It's a kind of read and If the hash fails then this means so either the data has been modified at the author system Or it's got somehow modified on the way However, we have a fail. So we will inform all the involved parties So the author has a chance maybe to to check what happened to the data and to see that everything is fine again And but we still have to somehow note this in the reputation calculation and There it's the next challenge. So how will the details of the reputation management work? So if you fail once probably this will not have an impact on your reputation, but if it fails continuously so then maybe the data is not there anymore, maybe the data has been changed or Maybe your research database is not really reliable, but this is has an impact on your reputation Yes, so to sum up What is very nice in the product we have in the end is somehow a combined top-down bottom-up Approach because we have now we're looking at the open science in general. It's more like so these are the process in place How can we implement them in the system? But on the other hand we come from the from the discipline from Plasmon research, so We come really from from the data from the researcher and from their requirements So we can fill the gaps in between these and iteratively. It's very nice and Yes, so the the basic process they are relatively straightforward. It is those that I have shown but The devil is in the details. So the additional security for instance So how can we push out the smart contract to do the hash actually locally? How can we have a meaningful reputation management that really will be a replacement for the metrics available today and Then how to implement a reliable peer review process, so which you are working on Yes, and we are in the process of joining the blocks book because we also think it's not There's no need to build another blockchain but if we can do our things on their network, it would be perfect and So I think I just got the message from James that we are in the consortium. So We're going ahead with that. Thanks Yeah, the devil is in the detail. I think a lot of Objects that tied to like that new new incentive structures suffering from these devils, right? Yeah, yeah, yeah, okay. So are there any questions? Yeah, okay Okay, thanks for this nice talk. I'm just wondering To make this really work if I understood it correctly and please correct me if I'm wrong You would he said this this would bear fruition if you have a lot of inter comparable Measurements, right? Let's say you were measuring a fundamental constant and everybody like like would would would give a certain value And you have a certain variation and those far outliers would would be something you'd Give a negative reputation or somehow downscore but often in most practical experiments It's essentially not only the data set, but unfortunately. I mean ideally it wouldn't but unfortunately they also depend on your Specific setup, right? So you have a certain experimental setup and Unfortunately the the result of that measurement is not or is dependent on how you measure it, right? Especially in life sciences and so on right so how many like like what is the typical cases where you would have a lot of Kind of things measure being the same thing measured from different groups with let's say Expecting the same results. So you have like a critical mass of measurements where it makes sense to really intercompare them and then validate them with your systems So we're not really comparing results to each other. So But we would like to to have experiments well documented so that they could be repeated in a better way and Also that if you find out that it's But that's not really in our system But if you see that these results are reliable when you so you only have to repeat maybe one experiment And then you can believe okay, you can trust the research data from these group because they Use the setup that they have in the metadata. So it's a bit more in this Are there any more questions? Yeah, okay, what is the status of implementation and is there any like Any implementation that can be test driven any time soon or what's the time frames for that? I ask it because I find it very very interesting Yes, so I mean there you can see we we are started more than one a year later than Artifact for instance, so it's not so much implemented We have implemented the verification process, but this will probably be replaced by blocks back and what we we have done Now we have defined those use cases So we could start with the implementation as soon as we have access to the blockchain or API So it it will we will probably take a small step. We were discussing with blocks back to have Some kind of collaboration of also in the direction of review process so we will see if something happens there in the next month is but To be honest, I don't think that we will have a whole system that would be comparable to Artifacts, we are more we are a research project and we would love to be in this direction But we by ourselves are too small. That's why we joined blocks perk and We will probably implement parts the parts that are most relevant for our project and then we will see how far it goes