 Good morning everyone. I'm Alex from catalysis. Good morning. I'm yours from digital science. First of all it's nice to speak to a crowd that is very positive about blockchain for change. So that's a good starter. We go to the first slide. So we want to talk about our blockchain for peer review initiative. It's an initiative we started about half a year ago, I think a bit longer, a bit more, together with catalysis, together with Springer Nature, Taylor and Francis, Cambridge University Press and Digital Science. And we want to talk today about progress, about where we are, what we aim to do with the initiative and also about lessons learned. But I want to start with some general observations about blockchain for science and I think where we are at in our efforts. And for me it's a good moment because last year, November, we published the blockchain for research report and it's really amazing to see how much progress there is in the last year. I think the first mentions of blockchain and the application to research and scholarly communication was a few years ago. It started with some blog posts. I think 2017 was really the year of ideas. Of course, Sunke was instrumental, very important with his open documents. And I think it's indication of, you know, your leadership that many of the ideas that you actually suggested and proposed with your group are actually being realized at the moment. So 2018 was really the idea at the year of concrete applications. So I haven't counted them but I think we are in a dozens of applications that's really encouraging. At the same time, I think we're going to the next phase, which is, and the former speaker Ava talked about it, is making sure it's actually used. And I think that might be even more challenging than thinking about it and building applications. So how can we move the ecosystem? How can we move researchers, etc., to use the blockchain? I think Albert Einstein said that politics is more difficult than physics. And I think it's also the case for computer science. So the big challenge really is going to be how can we make sure that this is going to be widely adopted in the industry. And I think there are a couple of things that are going to be crucial in that. First of all, of course, science and research didn't start yesterday. It's something that is hundreds of years old. And with that, history becomes legacy, not only in technology but also in processes, in culture and expectations. And we have to also think about how to change that. So what is going to be the catalyst for us? So what is going to move the ecosystem to a new way of working? Is it going to be government? Is it going to be funding? Or is it going to be researchers themselves? I think that's a very important question we have to answer. If it's about researcher, I actually liked your presentation very much. I think it's very important to make sure that they are on board. And of course, everybody is worried about openness, about transparency, about the metrics that are limited. But at the same time, of course, the short-term goal of researchers is reputation. And I think it's indeed, it might be surprising on the first hand that young researchers are not open to new tools. But I think it's very logical if you think about that they have to build their career. So they have to build their publications and citations. So it's very important that we address the reputation aspect in blockchain. The last remark I want to make is about open science. I think it's a no-brainer that science has to be more open, more transparent. I hear an echo. Is that just me? Do you still hear me? At the same, let me put it a bit further. At the same time, open science means different things. And we have to think, differentiate them and really think about what to focus here. Open science means sometimes not commercial. So getting the commercial companies out of the ecosystem. It can also mean open access, which is of course a separate thing. I mean, open access doesn't mean that the commercial companies are not going to be involved anymore. So what are we focusing on? At the same time, open access as a model also has its disadvantages, of course. I mean, I've been around in this industry for 20 years. The first talks about open access started 20 years ago, and we're still about 30% of the articles that are open access. Now, there are many reasons for that, but with open access also came some disadvantages. Pre-ordinary publishing, for example, or the challenges for researchers from the global south to get published. And I think we should be open actually to alternatives because the blockchain actually offers an alternative. That's micropayments. So let's also see whether that might be a more sustainable business model for the future. Open science also means open data, of course. And I think there, especially with blockchain, we have huge opportunities with tokenization and ensuring intellectual property. And sometimes it's also used in open peer review. And I think that's actually yet another topic because the reason, of course, there is single-blind and double-blind models. It's not because of commercial interest. It's because, of course, confidentiality and ensuring that researchers feel free to give candid feedback on articles. So without giving any recommendation, I think it's important that we distinguish those experts of open science. So let's talk about our initiative. What we try to do with our initiative is solve one challenging aspect of scholarly communication at the moment in its peer review. And there are actually many challenges. First of all, peer review is not recognized the way it should be. Again, researchers are mostly valued or assessed by a number of publications, citations, but there are activities that are equally important. And peer review, of course, is a very important one there. And yes, there are efforts. Definitely, there's publons. There's orchids, of course, but still it's not as recognized as other activities are. There's an increasing difficulty to find suitable reviewers. There are more and more publications. There are more and more publications from China. And they are making, you know, big developments in terms of publication, but they're not participating in the review process as much yet. So that puts a lot of strains on the existing reviewer pool. Unfortunately, there are cases of fraud manipulation leading to retraction. And basically, the peer review process is just simply too much of a black box. I think John Tennant said once we have to open the black box of peer review. And I think exactly that's what we need. We have to be more open, more transparent, and show more about what the process entails. Because, again, it leads to a lack of transparency and a lack of trust in science. And again, if you look at the problems around irreducibility and fraud and manipulation, it's really something we have to work on. Then the question is, of course, what do we do about it? So what our mission is, what we feel is that if we allow all the participants in the ecosystem to share peer review information, that's going to solve many of the problems I talk about. For example, when you talk about finding a reviewer, publishers are more or less dependent on their own database of reviewers. But of course, if you consider to select a reviewer for the review process, it's very good to know that the reviewer also did reviews for comparable journals at other publishers. It's also great to know that, hey, that reviewer, I might invite him or her, but he or she is actually involved in another peer review process at the moment. So it might not be wise to invite the reviewer at the moment. So we can coordinate the review process and make it more efficient by starting to share the information. Recognition also of course starts with having your data. If you don't know what the review process, who reviewed what, you can't recognize them. So if we collect on a structural way information about review process, and the reviewers, we can also properly recognize them. And of course, transparency. We have predatory publishers, we have review process which are not entirely incorrect. If we store information, again, the metadata about the review process, we create an auditable trail, so that we, in case of questions, in case of concerns, we can always go back to the blockchain and show actually the review to place or didn't take place. So that really adds to the transparency aspect of the peer review process. Very important aspect. We are building a private blockchain within the ecosystem. So all the participants in peer review, we invite to join us. So we have the purposes I mentioned already. We have Orchid, which joined us well. And also important, the funders. Because the review process for journals is now more or less separated from the review process from the funders equally important. But of course, there's a big overlap between those that review papers and those that review funds. So also combining that information will make the process more efficient and more transparent. So what did we do exactly and where we are? We are now, we build our proof of concept, which we delivered in September, according to planning, which is remarkable. And in a proof of concept, we're testing three things within the range of possibilities to validate the process so that we, using the blockchain, can independently validate the review process. So we don't have to rely on the publisher itself. Recognition, concretely, whenever a reviewer performs a peer review for the journals that are connected to the system, information about that is automatically sent to Orchid. So that adds to the recognition aspect and search. So already starting to use the system to make searching, finding, selecting the right reviewer, combining the information from multiple journals and multiple publishers more efficient. And for the more technical details, give the word to Alex. Thanks, Joris, for this intro. So I'm from catalysis. Catalysis is extremely young in this industry, only one year. So we have a very fresh view, maybe a bit naive. But what we would like to do as a company is to democratize the value of online contents. And to do that, we leverage blockchain technology. So our vision, and it's not actually contained to academia, it's more broad in terms of publishing, is really to tie together what we see as the three actors, the three main actors in this world of creating and consuming digital contents. So we have contributors who are interested in having their work attributed to them, compensated as well. You have consumers who want relevant contents, of course, and they want it to be available easily. They want it to be of a decent quality, and they want it at a fair price. And you also have people who check the quality of the contents, so reviewers. So we believe that they seek appreciation, sometimes compensation, and also being able to comment truly rather than just want to make sure that they comment it. So they want safety and being able to give their comments. So that's how we view how eventually the landscape should be. So how do we do this? So we're a tech company, so we build on top of a blockchain implementation using an Ethereum virtual machine, and we build various modules that basically provide specific functionality. So for instance, we have a micropayment modules, which allows you to get money from the existing banking system, so no cryptocurrencies, in and out. We have ways to identify content uniquely and register ownership. And various of the modules that you can then sort of tie together into a proper solution. So right now we have two solutions. We have a micropayment solutions. So for instance, should you have a blog and should you have loads of friends or fans who are willing to pay for what you write, we provide a WordPress plugin that allows you to effectively provide payments for your blogs. But that's more the commercial publishing space. In the peer review space, we basically take some of these modules and put them together to build what I will be discussing a bit more in a minute. So to do this, we stand on the shoulders of the giant. We don't reinvent everything. We do not build our own blockchain. There's just not enough time for this. So our starting point is the ethereal virtual machine, which is in the middle. However, we do not necessarily want to be tied to a lot of other great projects that are pushed on the blockchain. Therefore, we use a great blockchain technology, a tendermint, which for those of you who were there yesterday have heard a lot about. So we use this as our consensus engine and it exposes to us an ethereal virtual machine. And all of our top layer stack is Swift, which is a very nice language. So then what do we do with it? So we build our own libraries on top of the EVM. So we have a number of smart contracts, which we interact with using libraries that we've built. And because the libraries end up being in a server-side environment, we also have access to traditional technologies such as databases, payment systems. We can interface with mobile clients, web applications, you name it. The interesting aspect of using the EVM is that we're not tied just to tendermints as a consensus engine. So we can actually replace that bottom layer by a number of other blockchain limitations. Of course, there's Ethereum, there's Quorum from JP Morgan, there's Burrow from Monax and we use effemants. So let's go a bit deeper in how we build this. So we provide a parser, which extracts information from manuscript management systems, which are basically where the reviews are done. We process this data by chopping it up into various categories, the three categories that I will go over in the next slide. And we push some of that data in the blockchain, we push some of that data to ORCIDS, which interestingly enough helps a researcher to start to annotate their profiles with information that can be verified on a blockchain. So yesterday Rob had a question around, oh, what if I have 20 different profiles? Then the thing is once you start to accumulate information that pertains to your profile, in one profile, effectively it's not in your best interest to start to have 20 profiles because you don't link the profiles. So we start to give more value to a profile. And the way ORCID is done, of course, is not just us that can add value, other parties can add value, and then you can have a complete profile, which is more interesting. On the top we have a peer review query system. So that's an interesting part because that's the part that allows us to open up the data that you may not want to store on the blockchain to external parties. And this process resides very close to where the data is stored, which means that as a commercial party, for instance, you're able to selectively grant access to data that you have within your company based on very sort of sophisticated permissioning systems, which are stored on the blockchain, which means that you can use the best of both worlds as then you can share some data on the blockchain. But the things which you feel shouldn't be shared just publicly, you're also able to share it and control access to it. So taking a step back, how do we partition the data? So functionally we have three types of data. We have public data, which we're okay, and parties who want to work with our technology are okay to share publicly. So it's data that you should assume that you give it away. No one will, or everyone will be able to access it forever. It's a blockchain, after all. For us, that means relationship between entities. So for instance, relationship between an anonymized ID, which represents a reviewer, for instance, and a manuscript. It represents the states in which a review is. So for instance, if I've been invited to review a paper, it stores when I've been invited. It stores when I come back with my review. It doesn't store the review itself. However, it stores potentially ways to go back into the review information. So the second one is personal information. So of course, we will never store any data which can identify me or any reviewer on the blockchain, because then it's there forever. And then we have a problem with the European authorities around GDPR. We store information that allows people to claim that they are part of relationship. And then the last one is competitive information. So that's how I explain that our query engine allows you to expose data externally without giving away the control. So effectively, how this translates is that the blockchain stores references which are protected using a permission system back to internal systems. So the data flow. High level, the publisher provides us with data. So in this slide, we've mostly extracted the high level data that we store as an example. So what we get is a manuscript which is tied to a reviewer and which has a specific review states. We process it first so that we extract the information that allows us to identify people uniquely. And we have a mechanism whereby we can then hand over to an orchid profile, for instance. And if people do not have an orchid profile or do not want to take ownership of this information, that's also okay. But it means that the data that we will store in the blockchain is basically useless for anyone forever. And we annotates the data that then goes in the blockchain with this unique identifier. So we design this unique identifier such that it's not possible. So let's say if you review for two different manuscripts, you will have two different unique identifiers so that you cannot sort of try to do some statistical analysis on who has done what and come up with the identity of the person. So what I've described is basically a nodes in our blockchain system. So how would it work then? Then it's really multiple participants who are able to give access to their internal data to other participants. So the way we started, we were working with Spring and Nature, Taylor and Francis, Cambridge University Press, Cargur joined recently, and we're feeding our data to orchids. And really the blockchain is this connector in the middle which allows the information to flow between all the parties, but also to aggregate the information of all the parties. So then I've added a question mark which is anyone else who would be interested to add more data to the system. The little people on the sides are basically also the fact that while we do currently use a private blockchain, it's mostly because at the end of the day blockchain is a very, very young technology. And yesterday we talked about how actually it's difficult to provide a system that is perfect day one which means that you do not need to perform any upgrades or this and that. And we realize that we're not at the stage where we can say there will never be any upgrades. So we want to at least for the next couple of months keep enough control of that system so that we can go through multiple upgrade cycles and at one point be able to say with confidence, okay, we open it up and people can start to use it. And we're also aiming eventually to provide an API so that people can add more services on top of it. So now I'm going to go through an example of what you could do with this. As Joris said, in September we completed the first parts of the proof of concept which is mostly a backend system. But there's no UIs and without UIs no one's ever going to use it. So the next step for us is to start to implement user interfaces which will make it extremely easy and obvious as to what the benefits of such a system is. So let me take you through an example now. So let's say a publisher through their own internal systems have a certain view on a reviewer. So they can see that this professor Aldous Dumbledore has done a number of reviews for them. But it looks like this guy is actually not really interested in doing any reviews. If you look at some of the statistics, for instance, you see that the acceptance rate is relatively low and that most reviews are not delivered within expected times. You have no clue whether the guy is available or not. And you don't know if he's engaged in any other reviews than your own because effectively you have your own view. Now you go to another publisher, for instance, it's the same thing. They have their own view. And for this one, actually this guy is actually top notch, right? He's really delivering things on time. And all the time he accepts a review and they're very happy with him. A third publisher has yet a different view. And with the pattern I'm starting to expose there is that every party has access to part of the data. And what we aim to do with this blockchain is to allow people to share some of the data so that instead of doing statistics on a subgroup, you can start to do statistics on a bigger group. So when we tie all this information together, what you could get to is actually a much better picture of what reality is. So that's quite interesting. Now what if we expose this to third parties or more people who have access to this information and who can also add more information on that blockchain? For instance, what if this professor actually is extremely willing to state that not only right now they have nothing to do, so they're available for reviews, but also they're going to go on holiday in a couple of or they've gone on holiday already. And they are on holiday also now. But what if you could also push that information on the system? Then effectively, you can start seeing that you can continue to enhance the information and take this even further because you can give an even better view of what your current availability is. So that's one example of what could be done with such a system. And the back end is there. So we're seeding it with data more or less as we speak. And we're hoping to work on the UI sides in the next couple of months. And now I'll hand it over to Joris. Yeah, thank you. Last couple of slides. Maybe something about the planning. So again, it's very early days. We are more or less half year down the road. But yeah, this of course will be successful if all participants in the ecosystem contributed all journals basically share the data and funders and institutions, for example, start to use the data as well. So when we tested the proof of concept, then it would definitely for next year we will add more participants. We go to the MVP summer plan for the summer of 2019. And now we're discussing with the participants and what kind of functionality would like to see in there. And of course, indeed, it's going to be a mixture of making sure we have a robust backend, but also showing functionality that really triggers publishers, editors, but also researchers again to get to ensure that the adoption is higher. But again, this is all still something that has to be decided. I want to end with some lessons learned. A couple of things that really surprised us or positive or negative. One is Einstein said politics is more difficult than physics, but legal is even more difficult. Especially, of course, because of the GDPR implementation, legal teams are extremely cautious. And of course, we understand that. The complexity here, of course, is that we're talking about blockchain, it's a new technology. So, yeah, really, the delays we suffered are really partly due because of legal consideration GDPR and explaining what it entails and taking away the worries is really taking a lot of our efforts. The second is managing the hype. Hype can be a good thing because it means people talk about it, but it can also mean a negative thing. It can people, for example, are interested for the wrong reasons or they don't really understand what it is. So it's really make finding the balance of using the hype and sometimes steering away from it, for example, by stressing the advantages for the reviewers and the authors and not even talking about blockchain sometimes. Sometimes that just simply helps. And the third is, again, tied to that, yes, is patience and education. We are in an extremely early phase, again, only the last couple of years. And we should be really patient. We should not, I think, expect big changes next year or even the year after. But we should really steadily work towards improving the ecosystem in various ways with, I think, the long term in mind. Because if we rush too much, people might be disappointed. And the hype can really work against us. So with that, I would like to end. Thank you very much. And welcome any questions. So there's one question right next to me and the others who have a question just raise your hand and I'll come and get and bring the microphone. Thanks for the nice talk, very nice concept and idea. I would like to go a little back to the to the initial part that you said that you have kind of all planning to have a nonprofit organization is running this. And I think governance and the concept behind this would be very interesting because, you know, we as a scientific community, we had the issue that we basically relying too much on commercial entities and are kind of locked in this. And from this was like, I really like that you are going for nonprofit. What is what is the aim? How can people participate in their houses governed? Actually, this would be my my key questions here. Yeah, yeah, it's a very, very good question. You have scientists complaining about I don't want you have yet another system I have to work with. And equally, you have publishers that that say I don't want to have another non non for profit body that I have to work with. The good news is that in academic community, of course, we have fantastic organizations that already work on a non for profit basis to improve certain aspects of scholarly communication. Orchids, of course, is a fantastic example, but also Crossref. You have trade organizations like STM. So on the long term, it might very well be that that's the organization will be, you know, placed on the one of these. Again, we don't want to have yet another body. The good thing about Crossref, for example, is that like 4,000 publishers are already part of it. So that the confidence could be very well fitted there. But again, that's that's all for the future. But I definitely think that's that's an avenue we have to explore. Hi, morning. That's a really good presentation. I'm curious about the first thing you talked about where you said there are a lot of papers coming in from China and not enough reviewers. And I didn't really understand how your particular system addresses that problem. Yeah, it's it's it means that we have to stress more on coordination. It means that we have to give the editors the tools to pinpoint the right editors. It means that we have to make sure that reviewers are not inundated with review requests. So again, if we can make it more efficient, if we can give the reviewers a tool to say, I would not select that reviewer because he's not going to answer because he's doing two other reviews. Or hey, look at that reviewer. He's actually new in the field. And he, you know, he raises his hand. Please let me do a review for the journal. And we can basically make it more efficient. And therefore, again, degrees the review times. Yeah. Yeah. Yeah. Yeah, I think definitely another efforts on the way. I mean, the publishers are doing that themselves, for example. Again, we cannot solve. That's a very good point to make blockchain cannot solve all the problems. And definitely people recognize this. So this is this is definitely something that is that is being worked on. It's we don't see it as our responsibility. We just want to make sure that things are more efficient so that, you know, the pressure is managed in a more efficient way. Hi. Here in the bank. Oh, sorry. Yeah. So I just I just had land breakfast and I watched the presentation. I had to rush back to comment on that because I watched it in the live stream. I have to tell you, I'm friendly, but this is not blockchain for science. And I would even say that what you're doing doesn't even require a blockchain. Right. So it's basically a presentation of an engineering approach to to the problem of peer review that does not exist. And it's actually quite dystopian. You know, you're throwing around with terms like identity. You showed this example of having this Gryffindor person and then you might even want to tie in his holiday plans. I mean, I think it's great that Cambridge University wants to work with you. But I think that's a reason why then I wouldn't want to work there anymore, because it's just terrible. This is not how research works. And I mean, it's just a engineering approach rally writing on the on the blockchain hype. Okay. But it's not what really blockchain means in the context of knowledge creation. Thank you for your opinion. At the end of the day, and I think it came through prior talks, where I mean, if you as a scientist, you need to use tools. A lot of scientists are happy to build their tools. Others are not interested in building tools. The previous presentation was about taking open source tools and trying to fit them so that they could become something useful. This is also what we're trying to do. If you don't find it useful, that's absolutely fine. We believe it serves a purpose. Hopefully it will. Otherwise, someone else will come up with something else. But thanks for your opinion. So I'm here. Thanks, guys, for the presentation. I wanted to ask address this question, Alex, to you. You mentioned that the reason why you're using a private blockchain is because you want to test a lot of things before you go public. Why not use test.net then? One answer. CryptoKitties. Basically, and maybe, so I have a corporate background for a while, I don't like when I'm not able to control a system that allows me to guarantee a certain level of usability for the people who use it. The problem with testnet or Ethereum or any public blockchain is that because you have this original goal of being open to anyone who wants to use your system, it means that you're also accepting to share resources. And for certain things, it's OK. But I believe that if you want to be able to provide a system with any level of reliability, you need to have more control on what you have. To give you an example, I am not sure that CERN will allow anyone to use their systems, for instance, because they want to come up with their nice physics experiments simply because you need to be able to control what the systems you use to be able to get to proper measurements. So we could use a testnet, especially at the current stage because we don't have any scalability issues or anything yet. However, we would be exposed at the later stage to someone coming up with an amazing app such as CryptoGitties that then takes down the network and we don't have much to say to the people who use our system. Correct. Though again, I mean, the Ethereum community is doing an enormous amount of work to be able to scale their blockchain. Personally, as a technologist, I'm much more comfortable with the approach that Tenderman takes. And I would for now go for that. However, we've kept this compatibility layer, which allows us to switch back to Ethereum at one point if we get to the point where we think it's better. So there are actually many questions, but the time for the slot is over. Do we allow more questions? Yeah? OK, cool. So then you were? Could you please clarify what are the incentives for scientists to switch to your system in their articles and reviews? Thanks. Yeah, actually, I think it might very well be the case that scientists don't realize that they're using the blockchain because we're not replacing the existing systems. We're not placing submission systems. We're not replacing any of the tools they work. The only thing that what we want to do with the blockchain works in the background so that again, the selection of the reviewer is better. So the reviewer will just notice in the end that he gets more targeted invitations, that he can proactively indicate their interest and that the office will realize that the review times become smaller. So again, it's not a system that the researchers themselves will actively work with. It's again, the publishers and the institutions and things like orbit will interact with the system. Do you have any forms of appeal or accountability within your system? Why was the point of the open peer review and immutable record of it? What's the point of? Sorry, open peer review? Yeah, what's the point of the immutable record and the trail of the record if I'm asking you do you have any systems of appeal or accountability for the review process? So if I do a review and it's biased, what's, you know, how is your open blockchain immutable record solves the problem of the bias? Very good question. If you look at it in a different way, so right now we're building a system and we're opening it up to big publishers who have access to a lot of review data. And smaller publishers also. But the main point there is that we do not build a system that only big publishers can use. We hope that at one point anyone who is custodian of such data can see the data there. Then because it's an open system, it's a blockchain, it's immutable, it's public, we also hope that everyone can decide whether they feel that if they see, for instance, in an Orchid profile, that someone has done a review for a given journal from a given publisher, they may decide that they think, oh, these guys, I hate them and I don't like what they do, what they write, so I rate them as maybe 10% trustability. That's for everyone to decide. Maybe as an individual, you want sort of only to work with certain publishers who have specific ways of working that you believe in. In which case, it's all good. Our system doesn't preclude you from doing this. On the contrary, it allows you to basically be able to choose what source you want to use. And at the end of the day, maybe in 10, 20 years time, all the big publishers and the small publishers that we can't really know about, maybe they won't be there anymore. Maybe they will. But that's not for us to decide. The markets and you guys as researchers, as scientists, will decide. And if they're still there, then maybe the answer will be, at the end of the day, they do provide some service that is good enough. Maybe it's not perfect, but it's good enough. What we hope is that we provide a way to give choice. That's it. So as someone who's had a lot of papers rejected, one of the really frustrating parts of the scientific publishing process is that you go through the review process, your paper gets rejected, and then you start all over again with a different journal, with a different publisher. And also as a reviewer, what is also really frustrating is that you review a paper, you give constructive feedback, the paper gets rejected, and you might as well have not given all that constructive feedback. So what would be incredibly useful would be a system where the reviews can be passed between journals. And so you're not starting from scratch. And I know that some publishers do that, like within their own suite of journals, but doing that across would be really cool. Yeah, that's actually, we have two slides, Dex, and that's actually the fourth point on one of the other slide, Dex. Portability. And it's definitely one of the things we're aware of. So recognition, transparency, review coordination, and portability. Indeed, in the case of Cascades, that you can give permission to one publisher to the other, saying, please take over this review process, and here's access, for example, to review port. But also in case, for example, the ownership change of a journal, so that history can come along. So that's definitely on the agenda. All right, we are already running a bit late. Thank you very much for this talk. If you have a question, please catch Alex and Joris in the break. Thank you again.