Ties are uncomfortable and distract from clear thought. He appears a lot more intelligent and focused on the science by not wearing one. Just a thought. :)
Yep. I’m late to the party here but it’s more specifically taken from the top of the « Arc de Triomphe » looking down at the traffic on the roundabout around it. it’s called « place de l’étoile » or « place Charles de Gaulle » and is known for being a nightmare to drive around for any none Parisian driver (maybe even for a large number of Parisian drivers !)
It is congruently amazing to the topic that you are posting this online with the course materials on your website. Not many aspiring engineers and software designers get a chance to go to MIT and by making this open source you are providing a pathway not only for the future of AI and self-driving cars, but self-learning and the education revolution that is bound to happen. Thank you.
Maybe, ‘amazingly congruent’. Still, authoritarian language police may be correct at present, but the future always reveals them anachronistic and willfully blind. Language and words evolve over time and change meaning. Many great writers were surely physically abused by vicious grammar police in school.
It's so wonderful to be able to learn from you. Thanks to MIT for extending the opportunity. I'm thankful to be able to see so much information and learn from someone so educated. I'm really optimistic and interested to be part of my future because of technology.
3:04 - 😅 Kinesiology student. I know, I shouldn't be here, but I would happily sit in for Lex's lectures even though I have no idea what y'all are talking about. Sunglasses and a mustache are certainly the prerequisites for me to enroll.
I also like to remind myself that English isn’t Lex’s first language. I try to imagine myself leading a lecture on a cutting edge topic at a top university entirely in Russian
I understand your point but his comprehensibility doesn't really lose anything @ x1.25. ( May be because I'm already familiar with the topic but yeah. )
guys please help me on this - i had decided to choose the deep traffic and deep tesla as a part of my mini project well i wanna know whether after watching all the lectures in the playlist will i be in a position to solve the deep traffic project i want to know whether these lectures are enough to solve the deep traffic problem since i have just 8 days for my project deadline please reply
guys please help me on this - i had decided to choose the deep traffic and deep tesla as a part of my mini project well i wanna know whether after watching all the lectures in the playlist will i be in a position to solve the deep traffic project i want to know whether these lectures are enough to solve the deep traffic problem since i have just 8 days for my project deadline please reply
@m ・ ́ω・ guys please help me on this - i had decided to choose the deep traffic and deep tesla as a part of my mini project well i wanna know whether after watching all the lectures in the playlist will i be in a position to solve the deep traffic project i want to know whether these lectures are enough to solve the deep traffic problem since i have just 8 days for my project deadline please reply
I wanted to start a self driving car startup at the beginning of 2022, but I think what this talk have showed me is the level of information overload that we make a self driving car process before making a decision which is very ineffective.
Thanks for uploading these videos. Last semester in my university, I took two courses Neural Networks and Machine Learning but I left them in the middle. I am taking these courses again next semester and these videos would of great help and motivation for me.
This is such an up and coming field. Back when I was a student I watched that DARPA stuff live and it was considered the future and state of the art. Now it's history being laughed at. Also noting that the video is from 2017 which pre-dates GPT but has some future looking references to it, also now history. The future is coming fast, I wonder what is around the corner for AI.
Universality - for any arbitrary function f(x) there exists a neural network that closely approximate it for any input x Special purpose intelligence - bedrooms, sq feet, neighborhood -> final price estimates
At 48:36 you say that a neural network's number of nodes and layers is fixed. But what about the Dropout of Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever and Ruslan Salakhutdinov (University of Toronto 2014)? Doesn't that publication state the opposite? Thank you for the lecture :-)
Good point. I didn't make that very clear, so this is a good opportunity to clarify. For dropout, nodes and connections are temporarily removed during training to avoid overfitting (make the network more generalizable). The fundamental structure of the network remains the same, at least in the sense that new neurons are not formed as part of the optimization. The "growth" (aka learning) of the network is in values on the parameters and not the increase in the number of parameters. Artificial neural networks are inflexible relative to the human brain is the point I was trying to make. On the other hand, maybe our brain is indeed running a clever version of dropout...
guys please help me on this - i had decided to choose the deep traffic and deep tesla as a part of my mini project well i wanna know whether after watching all the lectures in the playlist will i be in a position to solve the deep traffic project i want to know whether these lectures are enough to solve the deep traffic problem since i have just 8 days for my project deadline please reply
42:10 is interesting. And worrying. Computational irreducibility of self driving cars. We won't know what cars would do in some circumstances until they do it.
wow that is what I would call learning 2.0. I never signed up for this, nor am I studying this field, but I love to follow it anyway. THANK YOU for sharing!
guys please help me on this - i had decided to choose the deep traffic and deep tesla as a part of my mini project well i wanna know whether after watching all the lectures in the playlist will i be in a position to solve the deep traffic project i want to know whether these lectures are enough to solve the deep traffic problem since i have just 8 days for my project deadline please reply
I think there’s a great future ahead for this fellah. Only criticism is that his clothes are far far too bright and colourful. Wish he’d tone it down a bit.
19:46 I think you mixed up Tokyo with Paris :) This is Place de l'Etoile, recognizable with the triangles shapes on the place which form a star from above !
priorities in the algorithm used for that red car at the beginning look to be: 1) maximize lane distance from other cars 2) keep a minimum distance to other cars ahead in lane 3?) stay to the right?
guys please help me on this - i had decided to choose the deep traffic and deep tesla as a part of my mini project well i wanna know whether after watching all the lectures in the playlist will i be in a position to solve the deep traffic project i want to know whether these lectures are enough to solve the deep traffic problem since i have just 8 days for my project deadline please reply
Take whatever the most powerful GPU can do graphically. Now, imagine that backwards. Matrix transformations on each layer depend on every other layer simultaneously. Apparently continuous values were only suggested this year as opposed to discrete.
Sir, I am a researcher from Pakistan working on self driving cars using deep learning. Your way of teaching is excellent and very effective. I need the presentations slides if you may share with me.
Love your videos sir, I have one question from this lecture: If all the cars in the future are self drive and they are connected of a same network that each car goes 65 mph then there shouldn't be any traffic... if all cars are connected to a ML smart system to connect all cars then they all can change speed based on any other car in the network so traffic minimizes... no need of traffic lights anymore!! 😆
41:20 I immediately thought of the 11% chance of survival little girl in "I, Robot"... sometimes humans have to make calls that involves self sacrifice, and each human will make the call based on their own psychology, I guess it all comes down to learning and randomness anyway, since not everyone will choose to sacrifice themselves to save another.
Looking at the figure at 27:26 causes me wonder why the green bar is not configured to calculate where/how to hit it based on the location of the white bar and it's probability of successfully hitting the ball back.
This principle, for example, would be like a car calculating if it's possible to make a lane change or maneuver while taking the abilities (not only location) of other cars into consideration.
Deep learning - pr term for neural networks. Network layers that have many layers More data we give Imagenet classification error Translation of text in images
My name is Lex Friedman, I work with an amazing team. I am friends with Joe Rogan, I do mixed martial arts and run a podcast. I expect all of you build a bad ass car and do ninja kicks.
This guy is so brilliant. He will be a great podcaster one day.
Lol
Ties are uncomfortable and distract from clear thought. He appears a lot more intelligent and focused on the science by not wearing one. Just a thought. :)
Haahaaa
Omg dude- you’ll be astatic to know he had one! Lex podcast!!! :)
Maybe get his Ukrainian American friend to help him teach....
Never thought I would be able to say I knew something that Lex did not. The traffic video in question is from Paris. Schooled.
At first I see, that's right hand traffic, not Tokyo / India. Then I thought, is that Arc de Triomphe?
(I watch roundabouts)
Yep. I’m late to the party here but it’s more specifically taken from the top of the « Arc de Triomphe » looking down at the traffic on the roundabout around it. it’s called « place de l’étoile » or « place Charles de Gaulle » and is known for being a nightmare to drive around for any none Parisian driver (maybe even for a large number of Parisian drivers !)
Also your parents’ name
Lol
19:15
It is congruently amazing to the topic that you are posting this online with the course materials on your website. Not many aspiring engineers and software designers get a chance to go to MIT and by making this open source you are providing a pathway not only for the future of AI and self-driving cars, but self-learning and the education revolution that is bound to happen. Thank you.
Lots of free resources can be found on some of the websites of these colleges. Found some awesome texts!
@@pearlrival3124 learn English.. lmao
@@pearlrival3124 what do you mean ? of course it's a word. Explain incongruent in that case
"congruently amazing" doesn't mean anything
Maybe, ‘amazingly congruent’. Still, authoritarian language police may be correct at present, but the future always reveals them anachronistic and willfully blind. Language and words evolve over time and change meaning. Many great writers were surely physically abused by vicious grammar police in school.
Some people heard Lex's beautiful lectures even before his podcast came into the highlight. Blessed souls.
What a time to be alive. Thanks so much for these.
I definitely learned a lot!
Ahh, another Two Minute Paper fan :)
@@Dayvit78 thought the same. "WHAT A TIME TO BE ALIVE!"
Just imagine, 2 more papers from now, Lex will be a podcaster!
I think it's the first time I see Lex without a tie.
This gent is a good lecturer -- explained intuitively, covering the important aspects well. Keep it up.
It's so wonderful to be able to learn from you. Thanks to MIT for extending the opportunity. I'm thankful to be able to see so much information and learn from someone so educated. I'm really optimistic and interested to be part of my future because of technology.
This is awesome lecture! I loved every bit of it. Thanks for open-sourcing this course.
Great work, it's awesome of you guys to share all of the resources online with the general public.
great work of what? he is the enemy of civilians he is acia agent pushing nwo agendas
@@SirPraiseSun bro do something with your life, like make a difference or something, your past comments are visible
@@SirPraiseSun wtf
3:04 - 😅 Kinesiology student. I know, I shouldn't be here, but I would happily sit in for Lex's lectures even though I have no idea what y'all are talking about. Sunglasses and a mustache are certainly the prerequisites for me to enroll.
19:40 the video is recorded in Paris, champs elysees, the view is from the Arc de triomphe
I also like to remind myself that English isn’t Lex’s first language. I try to imagine myself leading a lecture on a cutting edge topic at a top university entirely in Russian
You also have to keep in mind that you didn't start learning Russian as an adult. The brain is extremely receptive to new information before age 13
Im do glad I'm alive at a time where I get to consume knowledge from Lex Friedman podcasts. Thanks so much
I have a few things to add. Try these in the programming:
CRASH=BAD
DRIVE [ME] TO [PIZZA]=GOOD
AVOID CURB {RIMS} {PEOPLE FEET SMOOSH=NO}
I really love how you teach and present these complex topics. You make the ideas very easy to grasp and also it's just relatable
Thank you for putting this on youtube for free, it's astonishing really.
Pro tip: regular speed, watch it twice, take your time, pause it, contemplate, and study the contextual information. You're welcome.
I understand your point but his comprehensibility doesn't really lose anything @ x1.25. ( May be because I'm already familiar with the topic but yeah. )
lmao
guys please help me on this - i had decided to choose the deep traffic and deep tesla as a part of my mini project
well i wanna know whether after watching all the lectures in the playlist will i be in a position to solve the deep traffic project
i want to know whether these lectures are enough to solve the deep traffic problem since i have just 8 days for my project deadline
please reply
guys please help me on this - i had decided to choose the deep traffic and deep tesla as a part of my mini project well i wanna know whether after watching all the lectures in the playlist will i be in a position to solve the deep traffic project i want to know whether these lectures are enough to solve the deep traffic problem since i have just 8 days for my project deadline please reply
@m ・ ́ω・ guys please help me on this - i had decided to choose the deep traffic and deep tesla as a part of my mini project well i wanna know whether after watching all the lectures in the playlist will i be in a position to solve the deep traffic project i want to know whether these lectures are enough to solve the deep traffic problem since i have just 8 days for my project deadline please reply
Fantastic job on this video, this is the best explanation of ANN's that I've seen yet. Loved the examples and illustrations as well.
I wanted to start a self driving car startup at the beginning of 2022, but I think what this talk have showed me is the level of information overload that we make a self driving car process before making a decision which is very ineffective.
Thanks for uploading these videos. Last semester in my university, I took two courses Neural Networks and Machine Learning but I left them in the middle. I am taking these courses again next semester and these videos would of great help and motivation for me.
This person should start a podcast, there is a high chance that it will get successful.
16:32 lmao the Russian accent slip
Hahahahaha
"you get to listen to me for the majority of the lectures" well you got it right I listen to you for hundred of hours now
FYI the crazy roundabout traffic video is from the Arc de Triomphe in Paris.
I thought it is located in Barcelona
Is there not lanes "lines" painted on purpose?
"Place de l'Étoile" to be exact!
You haven't driven (or tested your self driving algorithm) until you've done a 360 on Place de l'Etoile
Chris Mecton g
This is such an up and coming field. Back when I was a student I watched that DARPA stuff live and it was considered the future and state of the art. Now it's history being laughed at.
Also noting that the video is from 2017 which pre-dates GPT but has some future looking references to it, also now history. The future is coming fast, I wonder what is around the corner for AI.
Universality - for any arbitrary function f(x) there exists a neural network that closely approximate it for any input x
Special purpose intelligence - bedrooms, sq feet, neighborhood -> final price estimates
Amazing video, excited to watch the rest of them!
Thanks for this lecture, looking forward for more lectures. Helps me in medical science a lot
This Lex guy should try doing a podcast. Lots of potential.
Truly generous of you to share these lectures. Btw, it felt different to see you in this attire and especially glasses. Wow!
At 48:36 you say that a neural network's number of nodes and layers is fixed.
But what about the Dropout of Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever and Ruslan Salakhutdinov (University of Toronto 2014)?
Doesn't that publication state the opposite?
Thank you for the lecture :-)
Good point. I didn't make that very clear, so this is a good opportunity to clarify. For dropout, nodes and connections are temporarily removed during training to avoid overfitting (make the network more generalizable). The fundamental structure of the network remains the same, at least in the sense that new neurons are not formed as part of the optimization. The "growth" (aka learning) of the network is in values on the parameters and not the increase in the number of parameters. Artificial neural networks are inflexible relative to the human brain is the point I was trying to make. On the other hand, maybe our brain is indeed running a clever version of dropout...
Thank you for your reply
guys please help me on this - i had decided to choose the deep traffic and deep tesla as a part of my mini project
well i wanna know whether after watching all the lectures in the playlist will i be in a position to solve the deep traffic project
i want to know whether these lectures are enough to solve the deep traffic problem since i have just 8 days for my project deadline
please reply
Excellent introduction into Deep Learning! Thank you so very much Prof. Fridman. Looking forward to the future lessons.
Amazing Video Lex. Interesting to see what you used to wear prior to your black and white suit uniform.
Love these lectures.
bookmarked 44:00 1:03:00
52:30
I dropped out of high school . MIT is like being a Jedi.
We should all be thankful this cost us nothing. Thank you Lex.
I was wondering where I have seen this guy, then I remembered Prison Break
42:10 is interesting. And worrying. Computational irreducibility of self driving cars. We won't know what cars would do in some circumstances until they do it.
Jesse Pinkman teaching Reinforcement Learning! He turned his life around.
Not a bad 5 years vision when looking at the tools currently out there.
Thank you for sharing wonderful lectures
wow that is what I would call learning 2.0. I never signed up for this, nor am I studying this field, but I love to follow it anyway. THANK YOU for sharing!
"Dark chocolate" was in Swedish
Thank you for uploading this. Great quality and well articulated points.
Lex красавчик я все твои видео пытаюсь смотреть и понимать ты мне дохуя научил спасибо большое тебе и спать и спасибо что поддерживает Израиль
@меня
This stuff is brilliant, I really love to listen to his lectures and podcasts.
I had no idea Lex taught ML. I love surprises like this!
FRIDMAN@MIT EDU IS SUCH A GOOD EMAIL HELL YEAH DUDE
Thank you for the lecture and I'm looking forward for the next one!
guys please help me on this - i had decided to choose the deep traffic and deep tesla as a part of my mini project
well i wanna know whether after watching all the lectures in the playlist will i be in a position to solve the deep traffic project
i want to know whether these lectures are enough to solve the deep traffic problem since i have just 8 days for my project deadline
please reply
@@money_wins_controls How did it go?
I think there’s a great future ahead for this fellah. Only criticism is that his clothes are far far too bright and colourful. Wish he’d tone it down a bit.
Man I could listen to him all day, I wish he had a podcast of some sort
Thanks for making this public and not keeping this private at MIT! Thank you very much! :D
Thanks!
MIT classes look like so much fun..
Making a playlist might help keep track of all the lectures. Thanks! :)
Goog point, here is a playlist: goo.gl/SLCb1y
Amazing! Free courses, Thank you Mr. Fridman.
19:46 I think you mixed up Tokyo with Paris :) This is Place de l'Etoile, recognizable with the triangles shapes on the place which form a star from above !
priorities in the algorithm used for that red car at the beginning look to be: 1) maximize lane distance from other cars 2) keep a minimum distance to other cars ahead in lane 3?) stay to the right?
There have been so many "winters" for AI, that you could almost train a neural network to predict them xD
if True: log_progress_winter()
Hehe, the humble beginnings sir Lex.
God bless.
I can't believe this is free..... Did I just watch a lecture from MIT?
I took this class in person. It was fun.
Thank you for amazing course
guys please help me on this - i had decided to choose the deep traffic and deep tesla as a part of my mini project
well i wanna know whether after watching all the lectures in the playlist will i be in a position to solve the deep traffic project
i want to know whether these lectures are enough to solve the deep traffic problem since i have just 8 days for my project deadline
please reply
@@money_wins_controls nyes
You're an amazing person Lex
Really good lecture! Thanks for sharing all this knowledge for free! It’s like Elon said all the education is available for free !
Great Lecture Lex, ty for sharing it!!
Lex Fridman using MIT students to develop a better self driving network than TESLA itself. I see youuuuuu
An incredibly complex solution in search of a problem
Have to watch this another day. Just wanted to give ya a thumbs up 👍. Keep killing it lex.
Awesome lecture. Thank you very much for the upload Lex Fridman!
He made a space in my mind
Thank you Lex what a great intro ... following the class till the end!
great job. thanks for explaining a complex subject in simple terms
Take whatever the most powerful GPU can do graphically. Now, imagine that backwards. Matrix transformations on each layer depend on every other layer simultaneously. Apparently continuous values were only suggested this year as opposed to discrete.
Sir, I am a researcher from Pakistan working on self driving cars using deep learning. Your way of teaching is excellent and very effective.
I need the presentations slides if you may share with me.
www.dropbox.com/s/wo8tdnifmtqmmrm/deep_learning_self_driving_cars_2017.pdf?dl=0
13:39... where I am, the robot... truer words were never spoken.
Thank you very much @Lex Fridman
Lex Friedman is either the most sentient robot ever made or a brilliant human who appears to be half asleep.
Love your videos sir, I have one question from this lecture: If all the cars in the future are self drive and they are connected of a same network that each car goes 65 mph then there shouldn't be any traffic... if all cars are connected to a ML smart system to connect all cars then they all can change speed based on any other car in the network so traffic minimizes... no need of traffic lights anymore!! 😆
41:20 I immediately thought of the 11% chance of survival little girl in "I, Robot"... sometimes humans have to make calls that involves self sacrifice, and each human will make the call based on their own psychology, I guess it all comes down to learning and randomness anyway, since not everyone will choose to sacrifice themselves to save another.
Lex, put all them lectures on the tube! we need knowledge, and it should be free
I read it "De-learning for self-driving-cars" and was hooked.
The original is also interesting.
Looking at the figure at 27:26 causes me wonder why the green bar is not configured to calculate where/how to hit it based on the location of the white bar and it's probability of successfully hitting the ball back.
This principle, for example, would be like a car calculating if it's possible to make a lane change or maneuver while taking the abilities (not only location) of other cars into consideration.
Recurrent neural networks - sequence to sequence - natural language
Image caption generation
Drones
Utterly fascinating! Thank you.
41:28 Why not just give it more reward for moving forward in the race? (in the boat game)
Pass a Turing test
3. Automated reasoning. Use stored information to answer questions and to draw new conclusions
its like seeing a different person
Deep learning - pr term for neural networks. Network layers that have many layers
More data we give
Imagenet classification error
Translation of text in images
This guy is great! He should make a podcast or something.
This guy should have his own podcast
lex is multi-talented
Thanks for making this available online!
I love the browser! It’s the most accessible space.
A robot teaching about self driving car. I'd love to attain the class.
The Chuck Norris of Deep Learning
But where is Bruce Lee
28:18 What determines the amount of time needed to train a computer for a given AI task?
very good lecture
thank lex fridman for sharing it .
My name is Lex Friedman, I work with an amazing team. I am friends with Joe Rogan, I do mixed martial arts and run a podcast. I expect all of you build a bad ass car and do ninja kicks.