Minutes of the SIPB Meeting of 2018-03-12 The meeting was called to order at 19:30 by alvareza. In attendance were Student keyholders: alvareza, slz, nchinda2, cela, bpchen, ikdc, mtheng, dzaefn Associate keyholders: '() Members: twang6, rsthomp, amigdal, elefthei, npfoss, incematt, gibsons, mcameron, sargentd, wmoses Guests: joshuaa, yu513 Administrivia: '() Project Reports: cela: We had a Hyades hackathon yesterday. Hyades continues to be hacked, in a good sense. We've sent the requests for servers, contacted the media lab who's in the process of acquiring those servers ... So [that's] out of our hands. cela: We're trying to get different subnets for the machine room. rsthomp: HELLO I'M CAPSLOCK! I'm doing publicity things. ... If you have a hackathon project or have information on hackathon projects (especially whats.mit.edu) you should come talk to me. slz: So whats is a command, but nobody knows how to use it because it's a command. ... cela: As a reminder, scripts is in need of more people to help to maintain. Currently, all the maintainers have graduated years ago, and there's not been people to help with the backlog. So if you want to know more about how [MIT computing] works, or how systems work, you should come talk to me. [You don't have to be highly competent]. bpchen: Great help to the community. People love setting up websites. ikdc: Most importantly, the anime club website is on scripts. cela: You can even make your own website on scripts! cela: If you come help with scripts, it does not mean that you have to fix all the problems with scripts. slz: It is literally SIPB's most successful project. npfoss: I have a project proposal. I'm proposing a large project to address a couple of issues, specifically machine learning and large datasets and stuff like that. For one, large datasets are really hard to get and hard to use. [tea and tornadoes example] Some of the most helpful things for AI are things like MNIST and ImageNet and all that, where everyone agrees to use the same standardized dataset, and people can compare their results. This only works if it's centralized in one place and it's in the same format. This has the added bonus that you only need to go to one place to get your data. [combining datasets on the same axes]. This is an ambitious project, but it's a reasonable first step to "hey this is a thing that we made" and you can use it and there's an open source license ... like CSAIL has a lot of interesting datasets that they've spent a lot of time [preparing] but after they're done it's just sitting there in a corner and MIT students don't know about them. So even if there's just an internal MIT tool it'll be really useful ... One of the most important things in innovation is the time between when you have an idea and the time when you validate it, [and this database makes this possible]. So I think this will be an excellent first step. Obviously this will take a lot of collaboration between CSAIL and SIPB because we don't have the resources to host all this on our own. But I've had a conversation with MIT IQ ... That's going to be a new thing, they're going to be looking for stuff to show for it. They're going to be happy about a student group like us [starting this]. ... We also need CSAIL to help because they are the experts, and they are the ones for which this would be most useful ... Ultimately we need all three things. We need MIT backing, we need SIPB -- we need lots of us interested enough in it to make it happen, and we need [CSAIL], because [they are the experts in this so it will help us make it easy to use]. And this has the potenial for becoming something much bigger. So if you have feelings about this project in either stage, either the internal stage or the larger at-scale stage, please share this. slz: So will we be [just collecting the data or doing computational things?] npfoss: Oh! So there's this thing coming up called Hyades, which offers MIT students free compute. And what goes best with lots of compute other than lots of data to compute with? twang6: So reading the email thread, have you [looked at the alternatives]? npfoss: [haven't really looked into dataverse] ... ikdc: I thought the main thing about this project is collaboration between universities? npfoss: I think CSAIL is a good representative of the machine learning community? I think the point of doing this with CSAIL first is to highlight it and figure out what's good and not good, and once we have it we can [open it to the wider public]. cela: So [how is this different from dataverse]? npfoss: I've only had a day to look at dataverse. [Haven't really looked into it yet, will talk to CSAIL about it.] wmoses: Having had to go through a lot of hoops to get to data, a lot of data involves human subjects, even getting people to fill up a form is using human subjects, and MIT has this thing called COUHES, which says that you have to encrypt the data, [limit who you share it with], [other things to do with data storage]. But this could potentially undermine [the idea of making the data open]. npfoss: As you said, a lot of data is intentionally not public, so such a project will have to incorporate various levels of security and accessibility. Even if half the data on this service is something that only various groups on CSAIL have access to, they could still benefit from it. ... Such data could also come up in a search, [and if you want access to it you can go to the group and maybe do a UROP in it]. wmoses: This proces is horrific. I've had to go through three months just to ask people a questionnaire. cela: I'm also worried that dealing with data will require us to [deal with compliance issues]. So we should be careful about what data we're sharing. slz: [like astronomy data] wmoses: Maybe when anyone uploads data they must check that it fulfils [certain requirements], because just having the data [could get us in trouble]. Another question is how would you do the policing, because you don't want someone to do a manual approval process. npfoss: I imagine that there will be some sort of community process, and [maybe some committee can moderate it]. twang6: One of my first thoughts is that we should get as many researchers involved as possible, because we don't want to put the data in a format that they can't use. rsthomp: ... [You might want to reach out to other research groups to make sure that you don't overconverge to CSAIL]. wmoses: It's fundamentally important to get a data format before [getting the data], because ... if you want to go through the time-intensive process of collecting the data and putting it into this format, you don't want to do this multiple times. npfoss: The first [phase] of this project is going to be talking to researchers a lot and finding the best way to [design] this. alvareza: So with respect to LibrePlanet, we've talked to the administration and we've gotten the relevant information that they wanted about student attendance, and to them the percentage of MIT students that attend the event did not seem high enough to warrant this being an event that is categorized as an event that is hosted primarily by a student group. (meeting enters closed session) (meetings exits closed session) cela: Yesterday I took a look at the server room because we have been given an offer by mitch, [...] and achernya ... So now we need to take servers from the old racks and move them to the new ones ... which sounds easy but it isn't because we need to keep them running while moving them. slz: Also can we get our broken KVM replaced please? cela: Yes, when I'm not 100% hosed. slz: Can we allocate [...] for the KVM? [maybe we should take some time to think about this first] cela: Also we should clean up the server room, get rid of boxes. [We should figure out which servers we still need.] slz: CPW event -- sledgehammer to servers! In front of the stud! Also we should remove that gigantic server thing that's possibly radioactive. alvareza: I will be appointing Billy to be in CPW things as well! Do you want to do the CPW things? wmoses: Yes. alvareza: We all thank Billy for the prospectives -- I mean the prefrosh. [Also Lef.] elefthei: I remember we used to have cool exhibits in Midway, like the VT-100. Is it still working? [Yes.] elefthei: So if you have cool things that you want to show off, you should come talk to me. npfoss: Also the vive! The vive was really popular. slz: Did you know that we also have Adobe Creative Suite? Other: ikdc: It's snow day tomorrow. cela: There appears to be very fun failures in Kerberos on the exchange servers, as seen on class SIPB. I was affected by this but apparently it was resolved for me but not for other people. It's mostly affecting crvft, I don't know if this is because crvft are the only people who know how to configure their Kerberos for exchange. wmoses: [shall we mention the time travel last weekend?] dzaefn: While we time travelled the hour, a certain particular clock did not time travel with us. cela: It's almost time to start the meeting! ikdc: Motion to fix the clock. alvareza: Motion passes. Other Other: bpchen: Spark is this weekend! If you want free food check out dormspam for how to volunteer. cela: Hair dye continues to be a great thing that mankind has invented. nchinda2: Two guys in East Campus has made this thing called radiance [where you can input music and it will output lighting effects]. cela: One of the parallels of East Campus had a very fun experience with the Fire Department. Where "fun" means that they were not allowed to enter for several hours even though there was no fire. ... Don't cover your fire alarms. bpchen: Google Code Jam registration is open! If you like programming contests, they have this system where you can submit code to their servers... you can even write code directly in the browser! ikdc: I just want to point out that you cannot make a steak in Random Hall without setting off the fire alarm. cela: CPW is sooner than you think! nchinda2: MIT Bitcoin Expo is also sooner than you think. Specifically this weekend. It's free in the sense that you pay $50 and then they refund it to you when you show up. bpchen: We found the world's first elementary knightship in Conway's Game of Life. [description of Game of Life] ... It's a thing that moves at a certain rate ... cela: I'm a cel, so should I be alive or dead right now? The meeting was adjourned at 20:32. Minutes taken and submitted by mtheng.