Brief Comment on Peer Review

hi everyone this is charles hoskinson broadcasting live from warm sunny colorado negative 10 outside not actually too warm but certainly sunny today just a brief comment i saw on reddit somebody doing some scraping around and commenting on peer review of our papers and this is a common mistake that academics from other domains make when they start looking at things in the computer science realm so if you are a physicist a mathematician a biologist there's usually a canonical set of journals for your domain that are considered to be very prestigious and it takes months to years to get published in those journals conference proceedings are typically tier two or tier three things where if you can't quite publish in a journal you go to a conference and usually those are discussion pieces and they're not as prestigious and in many cases they don't go through any rigor or peer review at all computer science is a completely different domain it's actually the opposite in computer science and i wanted to make this video for those academics out there that are coming into our space you do not make hiring and tenure decisions in cryptography or compsci based upon general publications they do exist there certainly are journals in computer science land but conferences in the cs world are generally much more rigorous and they're much more difficult to get papers into for example crypto eurocrypt and ccs are considered to be tier one conferences in the computer science world in the field of cryptography and information security and generally when you submit a paper it has to go through a double-blind peer review process there's a referee process and only about 10 to 15 percent of those papers are accepted it's considered very prestigious to get papers accepted and 10-year post-doc and other considerations are actually made based upon your publications in that domain if it was that journals were certainly prestigious in computer science land we would aspire to publish in them but it simply is not the case for the most part for the practicing computer scientists at least in the domains of cryptography and programming language theory and so forth it's not to say journals don't exist but it's just done differently computer science is a relatively new field of academia and it exists in the 20th century mostly especially the second half of the 20th century so it's not quite beholden to the same traditions that academia has before it's much faster pace generally these conferences are annual some cases bi-annual and there's a there's much lower latency between when you submit a paper when the paper actually gets accepted and you present that paper some cases in the mathematical world it can take two or three years for a result to eventually percolate to a point where it actually considers to be fully peer reviewed and this is just not the case computer science so i figured i'd make this real quickly because it comes up every now and then we see it floating around telegram or reddit and it's an honest mistake that most academics make if they happen to be from a different domain and not realizing that computer science does things backwards for most of the academic world and if you want to verify that talk to any cs department and see the pup the cvs of the tenured professors at those cs departments you'll see the vast majority of their work comes from conference not from general publications that said there are other currencies such as particularly novel inventions particularly novel books that they write and so forth and so it's a much more dynamic field and there's much closer industry cooperation in computer science land than you typically see in a pure maths department for example so it's not uncommon for many theoretical and applied computer scientists to also be doing industrial research and this is especially ai domain in the cryptography domain because that happens to be where the most interesting deployments of the technology happen to be but i figure i'd just make a quick video because i see this every now and then and it's important to point it out that is how it just works those are the rules and i'm very glad it works this way if it was the other way where you had to go to a journal and wait two or three years to publish we actually would not follow with cardano a full peer review system we would probably have light touch and move on but because the fact that conferences are frequent and the review time is significantly smaller usually a few weeks to a few months once you've submitted as opposed to years in certain cases that that's a huge benefit and allows us to have our cake and eat it too inside the industry and there's well more than 10 000 citations for our pool of work 95 papers and the vast majority of them have appeared at some conferences and the really prestigious ones in crypto land are crypto and eurocrypt and so forth and every year we usually have something appear there so we've certainly met the standard of the computer science world and we'll continue doing that what's going to get interesting is that our department is starting to get our division our research through starting to get more interdisciplinary and there we're starting to go into game theory and economics into law we're starting to even do mathematical research again these types of things and there there is a a process that is more journal centric and unfortunately that's a bit too high latency for an industrial research group so we're gonna have to figure out how do we blend that research in a way where we can still benefit from what we're benefiting from from computer science land in the more traditional academic domains and that's a question for 2021 and beyond and it's going to be a fun research question we might even have to create some journals or endow some journals to have a slightly faster turnaround time or create levels of verification or review above and beyond a pass fail of did you just appear in the journal or not so it's an interesting academic question where this is actually becoming a big issue is in the biological domain especially in medicine and virology covid for example there are thousands of papers being written about kohid and they're all just appearing as preprints on archives as opposed in traditional journals and a lot of cases even when they appear in the journal they're not going through the normal peer review process which could take months to years if not longer and so there's a lot of discussion of maybe science should work a little faster and maybe we should be a little more open to things coming through faster we do in the computer science world and it's an interesting discussion and i hope it has a positive outcome on the speed of things because simply the industry just moves too slowly for it to be practical for industrial products in the general sense so common confusion but we resolve confusions here thanks for listening