 Alright, why don't we go ahead and get started again. So, welcome back to the second day of the virtual event in the fall 2021 CNI member meeting and for the next bit of time we have an invited presentation. I'm particularly happy to have this and we thought it fit best in the virtual meeting because it's a very long way to come from the UK to talk about octopus. So instead we are keeping Alex up late over in the over where she is. Just by way of background on this, every now and again, somebody really tries to come up with a platform that's intended to change the way science is done, the way science is communicated. And we've seen a few of these over the years, one that that very much comes to my mind is is the Center for open science and the work that they've done, although I think you'll see that has a kind of a different flavor than this. The octopus is getting a lot of attention in the UK right now. The just has made a significant investment in it and when we did our joint meeting with the just over the summer, our virtual meeting. And only there was, there was buzz around this and how octopus might over time change the landscape in the UK. There are some similar conversations happening around the European open science cloud, for example, these kinds of discussions seem to be in the air now. And that when I started asking around in the States, I found that not a lot of people knew about this platform so I thought this would be a wonderful opportunity to get this on people's radar and I am just delighted that the project leader. Alexandra Freeman has agreed to be with us for this. I just want to thank my colleague Victoria Moody at Gisk, who also has been very helpful in making this session happen. And with that, let me just express a really warm welcome to you, Alex, and over to you. Thank you so much for that introduction and it's, it's a real delight and an honor to be able to speak to you today and I've been listening to the sessions today and last night, well last night for me as well and it's, it's been really really nice to hear about other projects which have the same sort of feel as I think octopus has and it's not that common for me to find projects that have the same kind of ethos behind them. So it's really nice to be here. And to give you a bit of background, I'm going to talk today about what octopus, what I want octopus to be and and how it would work and how it will work when we launch next year. It's less about the technical side but do put in questions into the chat at any point so that when I come to the end I can talk about some of the things that you want to talk about. And to give you some background to me, I was a television documentary maker until a few years ago so my background is, I was trained as a scientist. So I'm coming from a science perspective, but I was in the media. And towards the end of 2016 I ended up leading a research group in Cambridge University. So I came back to academia and saw the academic science community and the problems that it was facing with the publication system, maybe from a kind of external point of view. So that's where I'm coming from. And I'm going to talk a bit about how I saw the problems and how I see octopus potentially solving them. So I may not need to tell you about what's wrong with the current system but I'll tell you about the things that struck me and a few of the top lines. We all know this one, it takes way too long to publish anything. I mean, COVID has really shown this in very clear terms and preprints have sort of had to fill a gap where journals were too slow. But why should COVID patients be special? You know, any advance, any scientific advance is urgent for the people that it might benefit. So I think that, you know, we have to develop platforms that allow knowledge to be transferred more quickly. And then there's no good guides to the quality of publications. We use pretty poor, pretty proxy metrics, as I'm sure you know. And then there are all the problems of reproducibility and publication bias and the beautifully named questionable research practices. Which again, I won't go into, we know about them. And the problems of actually finding anything because you probably know more than anybody with the problems with discoverability and accessibility. And a lot of these problems then have a knock on effect on the research culture. You know, there's this sort of hierarchical structure based on the inequalities of access to this one single measure of success, the mythical high impact publication. So I see problems within the research culture itself. And I think that a lot of it comes down to the fact that academic researchers are judged on pretty much just journal articles. Even though there are lots of other outputs for scientific research or academic research in general, as we all know, and the fact that academics aren't the only ones doing research in the world. There are lots of non-academic institutions doing really good research and really important research, which it would be great to be sharing and connecting with. But they aren't published in the same kind of way in journal articles, because journal articles take a long time to write. And the problem with judging academic articles, judging academics or researchers on the basis of journal articles is what's important in journal articles. This is taken from the Guide to Authors for one of the top science journals. And I think it's really clear that what it is defining is the classic high impact publication. What they're looking to maximise is they're looking for a piece of writing that carries a focused message in order to gain impact through readership and through citations. And you can read other guides that talk about a streamlined narrative and making your point. And this might sound perfectly natural because it's all classic advice about how to write persuasive articles that are likely to change people's thinking or their practice. And it's what I used to do all the time when I worked in the media. But I do think this is a fundamental problem for research. I think there's a principal issue here that journals are being pulled in two different directions. And this is to do with that, what they're trying to maximise. Because disseminating useful findings to practitioners and researchers is a really important job. And it's probably the one that drives most of their revenue. It's the popular bits of journals, the news, the views, the narrative reviews. But it leads to that aims that we saw in that Guide to Authors for things to be attractive, to be readable. And that's driving things like word counts and publication bias and a lot of questionable research practices. Whilst the other thing they're trying to do at the same time, they're almost being forced to fulfil this job. It's a rather thankless task of being a kind of patent office for science, of establishing priority, a place to record full details of methods and data and everything that's been done whether it worked or not. And that's a really specialised kind of communication. And it kind of gets naturally sacrificed because the incentive structure that would support that kind of communication is very different from the incentive structure that would support dissemination type of work. There are two different kinds of writing. The first one can be written in the style of normal communications. The things we're used to are a focus message and a narrative. The other one shouldn't be because when you're writing about what you've done, you shouldn't be trying to persuade people. And I think that that sort of focus there is the heart of the issue that there are two jobs going on at once. And I think what we can do is split those two jobs. And that's what Oxford is. It's designed to be the new primary research record and take on that second job and leave journals to do the first one properly. And then we don't have the incentives of one, of incentives of good dissemination, bending the incentives of good research. So Oxford is designed to maximise access to primary research. So all of the things that we think of when we think of open access, but it's also designed to set a new incentive structure to support good science. So as well as the things like being free to read, free to publish, digital first, all of those brilliant things. And I would love to make it language agnostic so that you read and write in your own native language, ignoring what everybody else is reading and writing in. But I also want to break up the concept of a paper as the main unit of publication. Because the research process naturally comes in a series of steps one after the other. And each step requires quite different skills, quite different resources. And forcing people to get right to the end of this process before sharing any of it, I think causes a lot of problems. You're retrospectively forcing a linear narrative. And you're only able to publish when the data sort of supports the hypothesis. And you're trying to make a really good readable story. Those are the kinds of things that drive things like p-hacking and harking and publication bias. So if we have a primary research record that's free of the dissemination incentives, we can have a structure of smaller individual publications and each of these blocks is a publication that more closely matches the actual process. So each publication type is linked to the one before it in the chain. So in octopus, you can't publish anything that isn't linked to an existing publication. And it has to be a publication of a certain type so that you're building these branching trees. But you can publish all these smaller units. And that has a lot of implications for what gets shared then and indeed what gets done. And those of you who are counting the seven types of publication and thinking an octopus maybe should have eight, the eighth type of publication is a review, the peer review. And reviews in octopus can be published attached to any other kind of publication, of course, but they're treated and valued in the same way as any other type of publication. They are just one of the eight publication types because reviewing is a scientific skill like any other and by incentivizing it and treating it like any other, we can encourage good reviewing and the research culture needs good reviewing. So this is what octopus looks like at the moment. We've built a working prototype and across the top there you can see the branching structure of linked publications. And we've just not got this super stuff. We've now got a navigation from a research problem to all the publications that are linked and related to that problem. But we've also got on the right hand side here a rating system because that allows well rated publications to come to the top of searches, but it also has a lot of other functions. So funders, for instance, could look for well rated and well reviewed publications and offer to fund the next step. And they could perhaps even offer funding to several groups simultaneously to carry it out, increasing reproducibility. And here you can see red flagging, which allows any logged in user who can read and write can also flag something where they think there's a potential problem. That problem might be potential plagiarism or ethical issues or statistical problems. And those sort of red flags get marked so that the audience can see the potential problem and the authors get notified so they can potentially resolve that problem because every publication in octopus can also be reversioned. So in the light of reviews or in the light of any issues that are being raised, you can republish, you know, reversion your publication to improve it. And I think there are a lot of benefits to this kind of structure. I mean, first of all, we've got instant publications. So there's no, all publications are as soon as every author has agreed to it. And that means that researchers are getting sort of priority as they call it for their work as soon as they can. So that's fast the dissemination of knowledge and it means that people aren't holding on to everything worried that somebody's going to scoop them. We've got smaller author groups because we've got smaller publications. So that means it's more meritocratic. You can see who is doing what. So that also is about accountability. So, you know, if there's a problem with research, where does the accountability lie? But meritocracy is important, especially for specialists. You know, the world has changed since the paper was invented way back in probably the 18th century or maybe even further back than that. Not one person does everything these days. We've specialized, we've become professionalized within science and those professions and those specialists need to be able to publish their own work in their own right to get recognition for it. And my sort of favorite thing is the no need for a narrative. I do think that narratives, although they're brilliant to make things very readable and very engaging are actually kind of dangerous when you're trying to record exactly what was done. They're good for dissemination. They're not good for the primary research record. And it means that we can share much more stuff. You know, you can share just a hypothesis or a small data set, all these really valuable things that are pieces in the jigsaw of knowledge. We need people to be getting them all out there, not waiting until they can build a picture with all their own pieces. And then the rating system. People often get really worried when I mention a rating system because of course, you know, research is a multi-dimensional thing. It shouldn't just be a number of quality. And that's not what I mean by a rating system. What I'm trying to do with Octopus is to really get away from proxy metrics and to think, well, how can we actually rate what we really care about? What is really important? What is a good marker of quality? And so by setting predefined criteria on which people rate publication and having different criteria for different types of publications, so what makes a good hypothesis is not the same as what makes good data. It allows us as a research community to find what good work means. So I think that we're actually trying to get closer to more meaningful metrics through a rating system. But of course peer review is very important, as I said. And in Octopus, as I say, we've incentivized it by making it a publication in its own right. And it's easy to write a review. It's just like any other linked publication. But being able to read other people's critiques of publication, this post-publication peer review, I think it's really important to help people understand not only the strengths and weaknesses of the publication that is being reviewed, but also learning how to critique the work of others. That's a scientific or a research skill, just like any other. And so we need to incentivize that. So post-publication peer review, I think, has so many potential benefits. And it's, you know, being adopted by journals as well. But in Octopus, I think it allows that collaborative working because it allows authors to reversion a publication in the light of reviews that could potentially even invite a reviewer to be an author on the new version. So post-publication peer review, I think, is important. And the fact that it's a publication type equal to all others. So good reviewers are getting credit for their work. And the other thing that I want to encourage in Octopus is interdisciplinary thinking. Because although research work is linked into these branching chains below an individual research problem, Octopus is also designed to encourage what I call horizontal linking. So if you find a publication that you think will be really useful, so say a physics equation on fluid flow that actually helps explain flocking behavior in starlings, then you can make a horizontal link between those publications. And then others who follow that link can then rate its usefulness. And just like any other piece of work that someone does on Octopus, all of these making links and rating links get put onto the person's individual publicly available sort of CV site within Octopus. And so the fact that everything you do on Octopus is available to see means that hopefully it incentivizes those things. So this is my user profile on the prototype. And you can see I'm an incredibly prolific and well-rounded author because I'm publishing lots and lots of different types of publication. I mean, that's probably not really very realistic. Most people will probably be specializing in one or two. And these sorts of easily visible metrics will allow institutions or funders, for instance, to see what type of researcher each person is from what type of publications they specialize in. Are they an ideas person or a data collector or an analyst or a good reviewer? And what do their peers think of them? So I'm hoping that these kind of useful metrics provide an even greater incentive for researchers to get their work into Octopus quickly and in the best form possible because it's all visible here. So as I say, Octopus won't be a good read, but that's where journals then take over in the dissemination process. So taking the content from Octopus and editorializing it for their specific readership. So just thinking then, I've tried to design a system so that it has benefits for different stakeholders within the publishing system. So sorry, for researchers, it should be removing the barriers to access for publication worldwide, giving people instant priority, making it easy to publish. I mean, it takes far too long to write papers. And it's easier to find relevant work because under a research problem, you can find hopefully everything that's published to do with that topic more meritocratic and you're getting more stuff out there. But for institutions, they can more instantly and easily assess what type of researchers people are, and it's a better way to see what their peers think of them because you've got open peer review. For funders, I actually think that it can really change the way that funders think about funding research because they can look for well-rated publications, as I said, and offer funding for the next bit of the chain, which potentially could avoid loads of waste for grant application process, wasteful on both sides. And again, they can easily assess individual researchers. And for journals, well, I think it lets them actually become more focused on synthesizing work and doing what really is probably the job that they're doing brilliantly at the moment. It's the getting the stuff out there for other people whose practice might depend on it, so practicing doctors, engineers, whoever. So finally then, I think Octopus, its login system and ease of publication means that it can be used and useful for researchers, hopefully in a really broad range of institutions, not just academia and industry, but also charitable areas, places where research is currently much less widely shared by the academic publishing system. I've had people contacting me from national laboratories, the police force, charities, NGOs. Research goes on in all of these places, but publishing a journal paper is such a big task. But it's only done at the moment by the people whose careers depend on it, which is academics. Hopefully, Octopus can change that and allow the whole research culture to move towards something that's more, that's broader, more inclusive, less hierarchical, faster, more collaborative, and generally a happier place to be. So that's my kind of the raison d'etre for Octopus. The sort of practicalities of it are that, as you've seen, we've got this working prototype, and we now have funding from UKRI to move to a launch next year in partnership with JISC, who are helping with the build, and the UK reproducibility network, who are really important for us in building community around it, and making sure that we use a test all the time and that we're serving people's needs. And I really want to thank those who helped us get as far as the prototype stage as well. But at that point, I'm going to stop because I really want to talk about the things that you want to talk about. So I'm hoping that when I look at the chat in a minute, I've got some questions from you guys, because I really value your opinions and your thoughts about who I should be talking to to partner with, how we can make this successful, and just learning from your experience, because you've got loads. Thank you very much. Maybe while we get the chat questions going, Alex, I could just start with one or two. There's a lot of interesting things here I'd like to explore, but one in particular is how you think about octopus connecting up with the published journal literature. I mean, in one sense, for example, I can imagine that a lot of the questions about supplementary data and things like that that currently come up with journal articles might be replaced by just referring to some complex of octopus objects. But I'm not sure. I'm sure you've thought about this and I'm wondering if you could say a little bit about your thinking there. Yeah, no, it's an important point because octopus don't set out to replace data repositories or protocols.io or all these other places that are specialists in being places where people can put important content. In octopus, we have the ability when you publish, say, a data publication, you put in the DOI or the URL to the repository where the data actually is. So in a way, yeah, we're not replacing those, but we are kind of, I hope, getting rid of that problem of supplementary information in journals, in that basically the journal article is then just a story that points you at all the really interesting and important stuff that's hidden in all the supplementary places and which are sometimes just, you know, PDFs or documents that are really quite difficult to work with if it's things like data tables or whatever. So I'm hoping that we're bringing that kind of system into the more digital first thinking, but we don't aim to be a data repository in the first place. So yeah, that's important that we want to be able to work with all of the institutions and all of the platforms that currently do already serve these specialist tasks. But one other thing that maybe I should say is that at the moment, because, as I said, every new publication in octopus has to be linked to an existing one, we've of course got the problem that we can't really launch with a completely empty database. So what we're trying to do at the moment is to use all the open access publications that we can get hold of across. And I'm hoping that I can now talk to Jefferson about the Internet Archives work, but taking all of that that we can and using natural language processing to extract from there the research problems and then kind of from all of those research problem statements to kind of aggregate them and then cluster them so that we can pre populate the octopus database with at least a selection of research problems that will form the slightly ironic backbone to the octopus, even though Octopi don't have microphones. But that's what we're doing at the moment is all that work to try and extract research problems to seed the database with, which is quite interesting work, but also quite tricky. Let me invite people to put questions in the chat for Alex and while you're doing that I'll just ask one other. As you think about trying to do roll out, are you targeting a specific discipline or subdiscipline for your, your initial roll out and what can you say about that. Yeah, it's quite a difficult decision in a way because I mean one of the things about octopus is that it's supposed to be there for all STEM disciplines and actually, I only say stem because that's my background and actually the more I talk to people about wider research, the more I hear that actually octopus may may work for a lot more disciplines within social sciences and arts and humanities as well. There are some areas I mean it might make sense to target particular areas and there are some areas where it's possibly a more natural fit or they're already exploring things like this and you know the life sciences or the medical sciences or psychology they all naturally fit this template quite easily and they're exploring things like OSF and the sort of project based working where you have pre registration so you're already publishing clinical trial protocols these kinds of things. Octopus would be a natural next step for those areas, perhaps more easily than than some others. So although I don't want to say we're going to focus on and launch with say three particular disciplines, it may naturally go that way. But I want to, you know, I want to try and allow it to be available to everybody at launch. But we may find that there are some that come to it more quickly than others. I've got a stunned silence on the chat. Yes. I don't know whether everybody's just thinking, God, this is just such a terrible idea or that everybody's maybe gone to make a cup of tea. Well, I will make one last call for somebody very kind. Thank you. Thanks for coming around. So if you come up with ideas and you want to pop them in the chat. I'll try and answer them in the chat over the over the next half an hour or so. Well, thank you so much and I'm really looking forward to watching what happens as you start making this available. I think it's a really fascinating idea. And, you know, there are many things I'd be very curious about once you get some data. For example, you know, how many of each object type, you're getting relative to each other. One of the one of the challenges with the number of experiments like this has been getting people to put the time in for reviews, even if the reviews are first rate, you know, visible objects. So it will be very interesting to see how that plays. Well, as I often say is that the technical build is the least of the problems for this. And what I sort of got a bit frustrated with when I started sort of talking about the problems of the publishing system, which was not a novel conversation to have within science or academia, was people saying, well, there is no better option. And so it's like, well, there is a better option. And then people would say, oh, but it would never work. And I always just think, well, you don't know it won't work unless you try. So my first step is to, you know, have the idea of what might be a better system and then living well build it and then do all the work that needs to be done to let people know about it and make sure that the institutions that can help drive this forwards understand why we're trying to do it. Because it's got to be, you know, we're trying to do it because it's better for the research culture and for science and for society. That's why I said it's so lovely to hear you guys in your projects and talk about it within this community because so often I talk about it within communities where the first question is always what's your business model and how are you going to make money which is what we're here for. All right, well thank you again. And, you know, as as Alex says she'll be here for a little bit longer so please. If you want to reach out to her in the chat do so. We're going to go on break until I don't know when I'm sorry I should have done that till 350. So we're going to have about a 13 minute break. And we'll see you back on the other side of the break so stretch your legs. Have a cup of tea or coffee, and we'll see you then.