Nate Veldt -- PhD Defense

Nate Veldt's PhD thesis defense at Purdue University on April 16, 2019.
Title: Optimization Frameworks for Graph Clustering

[@optimizedpran1247]

KZbin algorithm: wanna watch some dude's PhD defense?

[@122mlb]

why not

[@abdAlmajedSaleh]

I watched "Codemiko" maybe is that's why this recommended to me.

[@devina8812]

thats it? lol??

[@jacoblopez8446]

I have no idea why this was recommended to me but great work!!

[@NitishSingh-vq2gt]

OMG! I had the exact same thought. Anyway, Congrats to Nate!! :)

[@basit876]

It's your turn

[@user-cf5oh2vh8q]

I received this video may due to I am applying PhD degree in Purdue University

[@jacoblopez8446]

@@user-cf5oh2vh8q Im just a freshmen but Im definitely planning on applying to grad school. However before that, I have to survive analysis first lololol

[@user-cf5oh2vh8q]

@@jacoblopez8446 good luck, cherish the time being undergraduate, study more skills, you will be awesome in the future XD

[@Li0nheart1810]

Watching this while procrastinating my bachelors thesis

[@m322_yt]

same lul, what's your topic?

[@alexho5769]

same hahaha

[@Exachad]

You won the KZbin algorithm recommendation lottery.

[@philmccavity]

His explanations are so clear and jargon-free, it all feels intuitive and flows so well, it's as if you could have thought of it yourself. Inspiring!

[@ponsko777]

I have no idea why KZbin has recommended me this on Christmas Eve, but I congratulate you for your hard work. We are all so proud of you . Merry Christmas y’all

[@hasaniqbal233]

@Lewis Pusey nice

[@StefanM.-]

Sitting in Vienna, Austria 11pm drinking beer and watching a guy defending his PhD at a random university I never heard of in my entire life. Nice one, KZbin! And good job, Nate (guess so)!

[@ZapOKill]

same... but 7 hours and sober again

[@DanHalper]

Hopefully Nate will appreciate your efforts in watching the video.

[@1566marcus]

Same

[@forloop7713]

Whats up in Vienna

[@jocabulous]

you never heard of purdue university?

[@johnwig285]

Apart from the results of his research, what is impressive is that he managed to deliver a PhD presentation in a way that caters to the "average joe". Some presentations consist of so many technical jargons that audience lose interest.

[@caio_c2]

I believe it is because you are presenting to the chair, which usually have specialists wanting to know details of what you are doing of new and not what is already know.

[@harshithvallabhaneni2524]

I'm from India and its 4am here right now. Congratulations and all the best to brother Nate!

[@indiangirl_in_30s]

I am in 2nd year of my PhD and this really brought a smile to my face.

[@MarcCastellsBallesta]

Since everyone else has already mentioned the surprise of getting this in our recommendations... You mixed "choose your own adventure" with your PhD presentation. You're a presentation Jedi.

[@TheEarean]

I don't know why this was recommended, probably because I am a PhD student at Purdue right now, and always talk about optimization. Great research though, I am glad that I listened! I wish I could learn about slide 63 as well!

[@Zintilyaspin]

Sounds like you know exactly why this was recommended lol

[@derbigpr500]

It's not you that listened, it's your phone that listened to you.

[@TheEarean]

@@Zintilyaspin sounds like it ;)

[@nishadmandlik2108]

Good job! Good speaker and the ideas were presented superbly.

[@farouk-hjabo]

Also got this randomly recommended by KZbin, but man, your work is very interesting and you delivered the presentation smoothly. Congratulations Nate Veldt/Jim Halpert.

[@philippheller9439]

No idea why I got this in my recommended, not even remotely connected to doing any PhD work, but nice work!

[@Yagamifyed]

Absolutely fantastic presentation, well done!

[@greebfewatani]

Interesting and well put presentation This is my first time to try to understand graph theory and it's importance in our life. You answered all of this in the first ten minutes Thank you

[@solsoul6449]

Why did I think dissertation defenses had to be 2-3 hours? He completed this in 40 minutes. Clear, concise, and to the point. Not bad!

[@deebadubbie]

Congratulations Dr. Veldt! Your Prof. made a really nice introduction. I can't imagine how I landed up here though.

[@paulbetts4984]

Not sure why this was recommended to me, but none the less, congratulations on a flawless defence and cordial presentation. I look forward to more research in the future!

[@dheerajkumarkondaparthi]

Great work Nate! Congratulations!

[@MinimalisticM]

Watching this while procrastinating my master thesis

[@fengxiong6994]

I hope that I could finish my phd as soon as possible. Congratulations.

[@kevinjeong5525]

Great presentation Nate!

[@NoneOfTheAbove123]

Sir, you have been blessed by KZbin algorithm.

[@UnlistedCube0]

This is the most wild recommendation on my page so far

[@themobylogy5670]

This is big for us nate. Congrats

[@ShaunakDe]

I really enjoyed your thesis.

[@theabominabletrollman5763]

Great job -Jim Halpert- Nate Veldt, I was able to understand quite a bit and I'm not great at math.

[@synerjagg3516]

The Abominable Trollman haha nice one

[@pavankumarreddy7888]

I could catchup up to some extent ... but I got lost in the middle.... but I kept on listening .. then I slept...nonetheless this is a great work... at last I realized one cannot understand others 4-5 years of work in 40 min

[@p07614010619]

Great work, we all are proud!

[@tb0nestk]

After watching Nate’s PhD Defense, I now am enthused, confident and ready to take defense of my own PhD. That “PhD”, as a Pizza Hut Delivery guy.

[@uchungnguyen1474]

tks you KZbin, tks you Nate!

[@DeAngeloYouKnow]

congrats homie i'm proud of you.

[@danielarista1352]

KZbin knowledge graph strategy meeting: "We need to bring in some talent on graph clustering optimization" . KZbin recommendation engine meeting: "Do the thing".

[@omarhu983]

great dissertation! made it so I could understand

[@DaisyZhangAI]

lol why did i get recommended this video? congrats!

[@ricardoshillyshally5012]

wow youtube recommendation seems to know everything about me i go to purdue and i just took a graph class

[@bereketyisehak5584]

yeah, believe me google knows more about you than even yourself

[@WayneLynch69]

HAHAHA...SO TRUE! Even what I've watched on TV...

[@eastwestcoastkid]

Congratulations!!

[@jacfgerfung3658]

KZbin algorithm gets me here. Maybe I've been searching if I should take PhD. :D good work mate

[@nickklosky8476]

I'm a high school senior but you seem pretty smart

[@aravindgitamuralee5708]

That was really a great intro!

[@charlostainsawshades6962]

I don't know why this was recommended to me. I guess we are the chosen ones to become enlightened

[@vsdev]

Thanks for the inspiration

[@Spidermanisborn]

great work dude!!

[@commander_storm]

awsome work

[@epicmoviestv4774]

Congrats for the work mate

[@ABC-jq7ve]

Why is this in my recommendeds lmfao. Happy 2021 everyone!

[@RaviShankar-jx8qe]

Watched some videos on optimization and graphs then got this in the recommendation. Anyway, nice work!

[@vralev]

This bisection-like approach for finding lambda is called ternary search FYI

[@Imm23743]

No idea why this was recommended to me either. But I love the irony of how the most popular video on this supposed music channel is a PHD Thesis Defense.

[@annulrsolformrkelse4023]

Great stuff! Do you still work in research?

[@TheTomar33]

Great presentation, what could you apply this type of information toward sales? Which grouping would you consider?

[@moslehmahamud9574]

Random recommendation to a CS student, but im enjoying it! ^^ good job YT algorithm!

[@ninjasuper8]

Boiler Up! What a great defense!

[@Diana-ux8of]

good job nate!

[@zzkzkzzkzkzk8945]

This is going to blow up isn't it...

[@ce7545]

Not my line of work or interest but great work Nate!! I can only dream of having enough neuronal clusters capable of a completing a PhD i. mathematics!

[@aturninthegameof...4584]

Very cool, from Western Australia, other side of the world

[@GhastlyDesigns]

Yeah me too. Super weird to see Margarate river come up

[@rahusphere]

Congrats!

[@jianzhang2637]

My Ph.D. defense in Mechanical Engineering took me a total of 2 hours, the committee members constantly interrupted me by asking me tough questions during that long period of time.

[@jeremybuckets]

I guarantee when Nate walked in the room that day he did not expect 100,000 people to listen to his defense.

[@safalyaghoshroy2405]

Is this recommend to you in Christmas???

[@Bardia323]

It's interesting cause I'm about to take a course in "graph representational models" and this was recommended by youtube's algorithm

[@jaritos675]

great work! can't wait till someone implements this in R or Python or Julia. or maybe i could.

[@sungjuyea4627]

Is this something to do with TDA? Sounds interesting!(But I don't know much about combinatorics, sadly...)

[@nateveldt874]

Thanks for stopping by my thesis defense. This channel is mostly guitar videos kzbin.info/www/bejne/hoSqfql-jsxsjas If you're interested in computer science and applied math videos, you might get more out of my research channel kzbin.info/door/GK9c99M3_CvPU9Wm8cshyw or academic webpage people.cam.cornell.edu/lnv22/

[@TheGxb2162]

good job man!

[@timoijas5316]

Let's go Nate

[@ridnap]

After reading the comments I feel kinda weird that I actually searched for this :D

[@md.akiduzzamanabir3815]

Congratulations

[@karlrivera5583]

Congratsss

[@karllaurenzdiamante3784]

Like the others, this just showed up on my KZbin recommendations, but I'm all for it. I imagine that this'd have tons of applications for AI developers.

[@first9_yt]

Ok I am student of IT and thinking of pursuing PhD in Future, Insha'Allah. But are all if you recommend this?

[@ChristianRoland7]

What did you get your PHD in?

[@hancholo3411]

Damn i get recommended this video for absolute no reason

[@flatulencetheunendingii5815]

nice vid, doc

[@NortheasternChina]

Watching this while writing my PhD thesis... Every time I see others' work I feel ashamed cuz my work is so worthless nobody should listen to my defense.

[@alfredemil8017]

😃😃

[@thomasc3120]

You can’t think like that man. I’m sure your work isn’t worthless, certainly not if you’re writing a thesis on it, have confidence!

[@NortheasternChina]

@@thomasc3120 Thank you for your kind words. Anyway I’m going to leave academia and look for jobs in the IT industry.

[@thomasc3120]

@@NortheasternChina whatever you do just know you can and I’m sure you’ll succeed

[@NortheasternChina]

@@thomasc3120 Thanks. I know I can't be a failure knowing all the effort my parents have put in to raise me.

[@leif1075]

Who wrote the music at the beginning?

[@dgodgo5250]

The KZbin algorithm is really weird today. Great work though.

[@jchan0912]

Transposition or transportation

[@Pph0n1x]

chose your own adventure

[@speckinthedark]

"...and I have convinced you that LambaCC is the best thing since sliced bread..." XD

[@Ohushi]

no idea why im here but good job!

[@0xggbrnr]

This was recommended by KZbin, so I'm assuming this means that the algorithms KZbin used analyze its transcript are suggesting that whatever was said here is actually super important while constructing a general idea of how societies are functioning. I wonder if the research defended here can be applied to a lot of real-world network data and be used to draw a really interesting conclusion.

[@slicphone]

Good job algorithm friend!

[@felwind14]

Most probably they are using right now this thesis on us suggesting the we are in a cluster

[@hsingpeikao]

Great research. Gong rats! Got my MS from Purdue in 1987. Great campus, full of great professors and smart students.

[@rigolol7451]

I was very surprised to be able to actually understand this cuz I just finished my discrete math class 😂

[@brandontouchet451]

"This 'b??l sh??' curve" *rewinds* "This 'b!!l sh!!' curve" *le what?!*

[@jonross1347]

"bowl-shaped"

[@syoushiro1837]

why did people dislike this?

[@tricky778]

I'm confused. I expected a PhD defense to be a detailed questioning by a panel of experts, based on the thesis. But this looks like a lecture to 2nd year undergraduates based on the introduction to the new knowledge. Is this just a part of the defense showing that the material can be taught in a contemporary classroom?

[@WayneLynch69]

NOW that you mention it.

[@johnwig285]

The presentation IS part of PhD defense. The questions come later, which isnt shown in the video.

[@gregfleishman8437]

Also got this random recommendation - but have worked in graph theory optimization before (more on applications side). This lambda CC framework seems derivative of classic results in min flow max cut algorithms c.f. Hochbaum: www.jstor.org/stable/pdf/23482083.pdf I appreciate the clear explanations here, and the effort to be creative, but my first impression is that this is not particularly novel. Maybe this is addressed in the papers?

[@nateveldt874]

Hi Greg, I'm familiar with that paper by Hochbaum; it's a great one, but really not at all the same thing as LambdaCC. While there are some tangential similarities (1. they are both frameworks for graph clustering, 2. the sparsest cut objective shows up somewhere in both), if you take a closer look you'll note that the key contributions and main details are quite different. Hochbaum's framework is exclusively focused on ratio-cut objectives (which form one or two clusters depending on how you look at it), and its centerpiece is a min-cut max-flow based algorithm for a poly-time solvable ratio-style objective. As highlighted in my talk, LambdaCC is an NP-hard correlation-clustering based framework that can form an arbitrary number of clusters, based on the parameter lambda. It unifies a number of objectives that were not previously seen to be related (modularity, sparsest cut, and cluster deletion--Hochbaum's work touches on sparsest cut but the other two are quite different), and finally LambdaCC comes with new approximation algorithms based on linear programming relax-and-round techniques, and novel techniques for how to set the parameter in practice. There are many more differences I could highlight, but suffice to say that details are in my papers. Feel free to check them out if you're interested! scholar.google.com/citations?user=6is0_9wAAAAJ&hl=en

[@tototrapsilo]

I've never seen PhD Defense before, so i wonder is it commonl to assume that the people hearing a PhD defense in graph theory don't know what graph and NP-hard mean? Anyway it's good lecture, i love it.

[@hasaniqbal233]

i heard phd defenses are usually more rigorous and the 'judges' (idk who they are) often stop you and ask you difficult questions. But, phds are usually unique research so they prly wouldnt know everything you worked on

[@DoubleBananer]

*claps* I don't know why I'm clapping but I am.

[@flavioaguiar3824]

Man, i no have idea why youtube recomended this video. I'm brazilian, professor of History, in High School kkk. But, congrats, great job.

[@surprisedpikachu3987]

I am doing IB and my internal assessment on graph theory. I think youtube is trying to kill me.

[@batuhandev4847]

I study game design but got recommendation randomly for random guy's PhD defense lmao. Okay maybe that's because I look resource for my data structures and algorithm course maybe?

[@goldenjay6936]

Thanks youtube

[@semajxocliw]

1:31 extremely audible water gulp

[@timoijas5316]

That's some Tim & Eric type of stuff right there