Pieter Abbeel - Really Quick Questions with a Berkeley Professor
ฝัง
- เผยแพร่เมื่อ 3 เม.ย. 2018
- Dr Pieter Abbeel got his PhD at Stanford University under the mentorship of Andrew Ng and went on to become a professor at UC Berkeley. He's worked at OpenAI, Willow Garage and now Embodied Intelligence.Drawing on recent advances in Deep Imitation Learning and Deep Reinforcement Learning, Embodied Intelligence is developing AI software that makes it easy to teach robots new, complex skills. I caught him after he gave a lecture at Nvidia's GTC Conference in San Jose and asked him some really quick questions about his life and his thought on Machine Learning. Enjoy!
Please Subscribe! And like. And comment. That's what keeps me going.
Want more education? Connect with me here:
Twitter: / sirajraval
Facebook: / sirajology
instagram: / sirajraval
More Learning resources:
people.eecs.berkeley.edu/~pab...
pabbeel?lang=en
scholar.google.com/citations?...
Join us in the Wizards Slack channel:
wizards.herokuapp.com/
And please support me on Patreon:
www.patreon.com/user?u=3191693
Signup for my newsletter for exciting updates in the field of AI:
goo.gl/FZzJ5w
Hit the Join button above to sign up to become a member of my channel for access to exclusive content! Join my AI community: chatgptschool.io/ Sign up for my AI Sports betting Bot, WagerGPT! (500 spots available):
www.wagergpt.co
Here is the list with all learning resources suggested, along with their links.
Enjoy!
1. Fast.ai [9:37]
www.fast.ai/
2. Andrew NG’s deeplearniNG course [9:43]
www.deeplearning.ai/
3. Andrej Karpathy’s Deep Learning course CS231n: Convolutional Neural Networks for Visual Recognition [9:46]
cs.stanford.edu/people/karpathy/
cs231n.github.io/
cs231n.stanford.edu/syllabus.html
cs231n.stanford.edu/
www.reddit.com/r/cs231n/
4. The Deep RL Bootcamp (26-27 August 2017, Berkeley CA) [9:50]
sites.google.com/view/deep-rl-bootcamp/lectures
github.com/pmlg/deep-rl-bootcamp
5. Sergey Levine’s CS 294: Deep Reinforcement Learning [9:54]
rll.berkeley.edu/deeprlcourse/
sortezcouverts, where does he mention deeplearningbook.org? I guess I missed it.
Siraj, once again, Thank You!
no, thank you!
Chill guys. I'll take those "thanks" if you don't want to.
this guy just literally took my comment that I posted the first hour and made it "nicer"... th-cam.com/video/KRFMM4duLHg/w-d-xo.html&lc=z23sitnzjry4cl0rw04t1aokguwaj2ny1b4qeejrwnkurk0h00410
Mazen, I wish I had done it, so my life would be easier. Instead, I spent like an hour watching the video again and again and asking google to end up with this list. What a fool I am! Cheers...
Elektra Kypridemou you are my hero!
still waiting for really quick questions with elon musk
elon musk is God
elon musk is overrated
Damn that'd be a hell of an episode
i am going to Google I/O soon i asked google we'll see
Or Jensen Huang !!
Very nice to see super intellectual people saying kindness more than a few times with sincerity.
Pieter Abbeel - one of the greats. Good catch, Siraj.
There's not enough exposure of STEM careers in the world on TH-cam. This video is great.
Does anyone could post the links to the resources he uses to keep up to day in Machine Learning?
Hello machine learning experts !
I have a time series classification problem (stock markets) and I was wondering if I should lag the variables to take into consideration past values. I have already normalised them but still have this doubt.
As far as I understand these algorithms, for each line (equivalent to a month in my case) the prediction and training is made in function of all the observation I have in each column.
In the case of random forest where it has to calculate the best threshold value at each node to make the best decision, does it take into consideration value of past line (past months, i.e. past values) in oder to make the best forecast ?
Many thanks in advance for helping me with that !
Great video, these interviews are very motivating, especially for someone like me who is trying to learn ML online, but is detached from any actual data science community.
Where can I find an online course on learning deep reinforcement?
Can any machine learning experts help me ? I was just wondering if for a classification with time series I have to lag my predictors in order to take into consideration of what happened before ?
I was waiting for another one of this.
Hi. Can you provide links for his papers
Man i love this professor and his lecture ❤️
Thanks Siraj for making such inspirational interview video :)
Thumbs up for your videos! Great idea for the quick questions and definitely very inspiring!
What a great interview!
Please make it more helpful by adding the resources mentioned in the interview. Some of them are not clear.
Thanks Siraj.
Can someone list the resources for learning AI that Peter mentioned?
Valuable. Much appreciated!
Google I/O or WWDC,
-> NIPS and ICLR 🔥
Can't get any better than this !
Seriously !
boss
How what amazing all your videos, I like & love it, makes alive by listening, watching absorting all the machine learning topics, I am now so obseesed in Machine Learning....
Thanks Siraj, very helpful video.
outstanding questions siraj! boss
Thank you Siraj 💕💕💕 It is a very meaningful video！！
Great video thank you!
6:52 where can I find the published documents on oneshot visual limitation?? help
no worries ..found it => arxiv.org/pdf/1802.01557.pdf
Which course is the under carpasis course? I really didn't understand him.
Andrej Karpathy's cs231 course
Wow, those big question are making him tired lol Lots of great answers too 😁👏
How did you connect so smoothly with your professor? How did you build this friendly relation ? When I am talking with most of my teachers, I have that barrier. Please give insight. Thanks !
What are the resources he mentioned for ML and AI?
salehundam preetham I was wondering that too
fast.ai, deeplearning.ai and deeplearningbook.org
sortezcouverts thanks
sortezcouverts Thank you
Sir you are great
I knew you would do it again.
I like him. He reminds me a lot of me.
Nice.
Lol, if a CS professor can't keep up with the pace of research and mountain of papers, it's basically impossible for working programmers. Not that you shouldn't try to keep up with research, but it provides good perspective.
Hi Siraj!
Why isn't this titled quick questions with Pieter Abeel?
Hay Siraj, Could you ask Chelsea Finn all these questions please? My research is following on from hers. P.s. your videos are awesome!
why would you make an interview in such a crowded place? as a result we have too much unnecessary stress and bad audio.
how to catch a professor ?
Mac OS?!?
Great video Siraj👍👍👍👍... Can't wait for "Really Quick Questions With Elon Musk"
Can you make a video about future of reward function in RL?
Haha Shawshank Redemption is my favorite as well
Thats the best answer for why do you think programming is fun 5:08.
I like how he wants rnn to do the marking just as bad as students want nn to do their homework
wow man you should had a meet up at UCB lol
why do i feel like he is the real life sheldon cooper
He isn't robot like Sheldon cooper. He seems sensible and touchy person.
And Sheldon do not do any exercise or play sports except for paintball
Spirit animal: Koala.
I like this guy.
💓💓💓💓❤️❤️❤️❤️💓💓💓
What website do you spend the most time on..... We all know why that took so long ;-)
definitely robotsgonewild.com, that one vid with the grease zerks and the hydraulic oil.... crayzeeeeee
Who else were expecting favourite movie to be The Matrix?
So the chain of "Matrix" has been broken.
Based on Pieter's answers, I think I'm Pieter.
P.S. I mean the awkward part of him
"subleem"
Comment
subleeeeem
Why always 67 ??????
Go bears!
We need data but he won't give out his data to Google home or Alexa.
The thumbnail is horrifying xD
AHHHHM.
Look, that's Professer Johnny Sins!
lol so boring its like pass pass pass pass!!!!!!!!!
Bald Michael Fassbender
7:52 Easy one: buy bitcoin
3:26 Thats the worst thing you can do to a child. With homework you steal the time that a child needed for developing. How about get rid of homework and do research that helps people like cancer research
because some stuff, and some people need repetitions to learn, good homework is good, bad homework is bad. Also the teacher needs feedback to see if the student is assimilating the information, and homework is that feedback loop before a test, when it is too late to make adjustments.
Nice.