 man it's it's been a while obviously I don't want this video to be one of those you know oh sorry that I haven't posted in however many months because obviously I think you guys have a pretty good inclination of what has caused me not to be posting but we're coming up on the end of the semester the end of my last semester at Yale and in fact in five days I will have my last day of class and will officially be done with classes at Yale obviously I spent I spent this past weekend doing stuff with my secret society super busy didn't even think about school now I'm back and a lot of deadlines ahead of me to be completely honest it like doesn't feel real at all and I was talking about this with some friends the other day I totally just kicked the tripod it's almost like we said goodbye to Yale last March when all the COVID stuff happened and we knew we weren't gonna be coming back to finish like my junior year so like I I feel like I should be sad about graduation but like it's like I don't it's like almost like I've already like grieved the loss or the end of my college experience you know but it does feel weird to be like entering the world and becoming an adult and I can't even explain to you how burnt out I am of academic work from the semester especially since we didn't yell didn't have a spring break this semester I feel like that was a somewhat common thing among a lot of schools but rather they chose to give us you know random break days throughout the semester where we were like supposed to not be working but in reality like all of us were working because we had to catch up on stuff so it's been a long long haul and I had one week week left to finish my thesis to finish some final papers it's definitely very stressful but it's also like I I feel like I shouldn't be stressing about it because it's like my last week of classes obviously I have a couple more weeks till graduation still but it's my last week of classes and it's like why am I stressing about something that I've come to love so much so I'm just you know trying to not like it's not like I'm throwing my grades away or anything but I'm you know trying to come to terms with the fact that I'm leaving this place so soon and trying to you know spend as much time with the people that I love here friends professors advisors mentors everyone along the lines before I eventually depart but anyways that was a little sappy note I need to I'll update you guys on where I'm out with my thesis and some other stuff take you along for the rest of the evening but I think my rice is done my favorite things to cook since you know living off campus and going off the meal plan it's basically like a massive burrito bowl I just fill it up with a bunch of different stuff so with that I'm going to probably watch some Netflix eat this and then I'll show you guys how my thesis is coming okay dinner was great but now I'd love to show you guys kind of some updates on my thesis if you haven't seen my thesis video I highly recommend checking it out but I'm writing my thesis on the applications of deep learning to password cracking and unfortunately it's due in two weeks so I'm it's a rush to the finish line right now but I'm trying to have a solid draft completely done by the end of this week so I have five days most of my code is done at this point now I'm just retraining models with a you know more cleaned and concise data set but here I'll kind of walk you through what I've been working on so far so you can see here is my whole vs code terminal and you can see that I have three main models that I'm working with right now the GAN the LSTM recurrent neural network and then the V variational auto encoder the VAE so those are the three models that I'm working with right now it's been quite an interesting time I was able to get the recurrent neural net the LSTM up set up pretty quickly within the first couple weeks my thesis the GAN has been a constant fight if any of you or have any experience training GANs you know it's super hard dealing with mode collapse and gradient explosion vanishing gradients all those certain things so it's been quite the struggle that took like about six weeks and then I have spent the past two weeks I'm doing the variational auto encoder which is slightly simpler than the GAN that's still more complex than the in the LSTM network so everything's trained right now and then it's really interesting let me set this back up on the tripod but basically I was able to get access to the Yale high-performance computing center which is basically a big building full of supercomputers which because like initially when I was you know estimating how long it would take to train my first network just the simple LSTM one that that estimate ended up being like 60 days for a single epic so a single run through the data set which is 60 days on my normal computer which obviously isn't optimized for deep learning at all so after reaching out to my advisor and getting some help I was able to get an account on the Yale high-performance computing center which is super cool it did take some time to learn how to use because rather than just coding on specific nodes you have to like enter a specific interactive mode if you actually want to test stuff live because you can't test it on the node that you log in on like the actual computer and then any long-term jobs which most of my jobs do have to be long-term because they take you know a day anywhere between a day and like five days then I need to actually submit what's called like a slurm job so you can see an example of a slurm job here it's pretty much just a bash script that says exactly what commands I want to run and with the exact I need to also like specify the exact hardware that I want to use so then I just submit that job to a queue and it sits there and normally after 10 or 15 minutes it's accepted and it starts running and then I specify an output file for all of the you know everything that's going to the standard output of the computer so I can you know check up on that live so you know that's been kind of the process I'll submit a job and I'll sit here and I'll wait check in the next night see how it's doing see if it's training correctly make sure everything's working good if not you know cancel it submit the job again otherwise I just wait till it's done now the other side of my thesis is the writing aspect which isn't super intense it just involves doing a full write up for everything that I coded basically you know like a typical you know computer science paper so this is what I have currently a beautiful latex document I'm actually typing everything up in a Google Doc first so I can have all that spell check so I don't have to worry about that when I actually plug it into latex that's normally a pain but I'm about 10 pages in maybe 12 if you count the references and it's not like it has a specific page limit or anything it's more just you know how long does it take to actually properly explain your project so mine is probably gonna be somewhere in the range of like 25 to 30 pages it could be shorter but you know I thinned down the margins just the font so it looks super good but that is gonna be my main task for the next week so with that I'm gonna get back to work so it's almost midnight so I think I'm gonna call it a night because I still have to take the time to edit this video it's a very productive day it made a good amount of progress most of my code is pretty much where I needed to be right now there's a couple small adjustments I need to make otherwise I'm just waiting on these jobs to finish from the high performance computing center once those get back I'll be able to you know actually run some experiments have these things generate some passwords for me and see how accurate they are with that I really want to emphasize that this was very much just a casual like me sitting down filming and talking sort of video and that's very couple reasons one these types of videos I feel like I can get a lot more information and share a lot more with you guys like personally and get it up to you as fast as possible because it doesn't take as long to edit don't have to add any you know fancy transitions and all those sort of things into it as well although I do like doing that and I would like to put those in future videos just want to emphasize that this is kind of a different style that I'm experimenting with and if you'd like it you know drop a like down below tell me what you think in the comment section and if this is first video you've seen for some reason if you want to learn more about Yale me computer science anything along those lines feel free to subscribe so with that I'd love to continue sharing with you guys some of my life as I live out my final days at Yale and move on into the future hopefully I'll be back again with another video soon but until then see ya