Andrew Ng: Deep Learning, Education, and Real-World AI | Lex Fridman Podcast #73
HTML-код
- Опубликовано: 12 май 2024
- Andrew Ng is one of the most impactful educators, researchers, innovators, and leaders in artificial intelligence and technology space in general. He co-founded Coursera and Google Brain, launched deeplearning.ai, Landing.ai, and the AI fund, and was the Chief Scientist at Baidu. As a Stanford professor, and with Coursera and deeplearning.ai, he has helped educate and inspire millions of students including me.
This episode is presented by Cash App. Download it & use code "LexPodcast":
Cash App (App Store): apple.co/2sPrUHe
Cash App (Google Play): bit.ly/2MlvP5w
PODCAST INFO:
Podcast website:
lexfridman.com/podcast
Apple Podcasts:
apple.co/2lwqZIr
Spotify:
spoti.fi/2nEwCF8
RSS:
lexfridman.com/feed/podcast/
Full episodes playlist:
• Lex Fridman Podcast
Clips playlist:
• Lex Fridman Podcast Clips
EPISODE LINKS:
Andrew Twitter: / andrewyng
Andrew Facebook: / andrew.ng.96
Andrew LinkedIn: / andrewyng
deeplearning.ai: www.deeplearning.ai
landing.ai: landing.ai
AI Fund: aifund.ai/
AI for Everyone: www.coursera.org/learn/ai-for...
The Batch newsletter: www.deeplearning.ai/thebatch/
OUTLINE:
0:00 - Introduction
2:23 - First few steps in AI
5:05 - Early days of online education
16:07 - Teaching on a whiteboard
17:46 - Pieter Abbeel and early research at Stanford
23:17 - Early days of deep learning
32:55 - Quick preview: deeplearning.ai, landing.ai, and AI fund
33:23 - deeplearning.ai: how to get started in deep learning
45:55 - Unsupervised learning
49:40 - deeplearning.ai (continued)
56:12 - Career in deep learning
58:56 - Should you get a PhD?
1:03:28 - AI fund - building startups
1:11:14 - Landing.ai - growing AI efforts in established companies
1:20:44 - Artificial general intelligence
CONNECT:
- Subscribe to this RUclips channel
- Twitter: / lexfridman
- LinkedIn: / lexfridman
- Facebook: / lexfridmanpage
- Instagram: / lexfridman
- Medium: / lexfridman
- Support on Patreon: / lexfridman 
Наука
I really enjoyed this conversation with Andrew. Here's the outline:
0:00 - Introduction
2:23 - First few steps in AI
5:05 - Early days of online education
16:07 - Teaching on a whiteboard
17:46 - Pieter Abbeel and early research at Stanford
23:17 - Early days of deep learning
32:55 - Quick preview: deeplearning.ai, landing.ai, and AI fund
33:23 - deeplearning.ai: how to get started in deep learning
45:55 - Unsupervised learning
49:40 - deeplearning.ai (continued)
56:12 - Career in deep learning
58:56 - Should you get a PhD?
1:03:28 - AI fund - building startups
1:11:14 - Landing.ai - growing AI efforts in established companies
1:20:44 - Artificial general intelligence
Thank you very much
The reason why hand writing is fundamentally different and more powerful because you are in the act of recalling and creating neural pathways. It's like drawing you are creating something from nothninness
About time! ( 😂)
I think you forgot:
1:22:02 How to throw shade at Elon Musk without mentioning his name.
Jk, thanks for another great interview!
Absolutely loved this one!
I'm just a tradesman painting houses for a living but I stream these discussions in my headphones all day long while working. Even a luddite like me can gain so much from a guy like Andrew Ng - it's a quasi-spiritual process contemplating AI and how it relates to how I think and see the world. Even those not educated in this area ought to take an interest as its so intellectually enriching.
you’re an inspiration! everyone has humble beginnings
I do construction and I do the same thing lol. Cool to see I'm not the only one.
The work you do for money and your knowledge should not have to match. They are two different things. I am a software engineer and I want to learn to paint house. Because, that is a skill and activates different parts of brain. Education is everyones right. And, they should have access to it.
@@refactoredcodes Well said. Everyone should try to learn as much as possible. And like you said 'job's status level' should not define people. There are smart people everywhere.
Bunch of good will huntings over here
I have a PhD in AI and I can say that Andrew is THE BEST machine learning/AI educator I have ever seen on the field! The deeplearning specialization is just invaluable for any interested learner in DL.
Hey, is it a Ph.D in standalone AI or is it Ph.D in CS with a concentration in AI? (PhD AI or PhD CS with AI concentration).
@@pranav3658 Hi, there is no such a PhD in AI as the specialization. Usually, it would be PhD in CS, Stats, or EE with a focus on AI and machine learning.
Sherif Mosaad aah, thanks a lot for the reply! Also, I have an admit into the masters program in AI @ Boston University, AI is definitely what iam passionate about and a Ph.D is definitely a part of the bigger picture for me. Would you recommend I take it or rather settle for a more traditional MS in CS with focus on AI?
@@prof_shixo hey I am 17 still in high school. I did some courses in machine learning and some data science courses.Topics like Neural Networks intimidate me. I wanna ask which major should I pick as my undergraduate degree. CS, electrical and electronics, or economics degree?
@@shravani9106 CS
Whenever someone uses the word "It turns out", I remember Andrew Ng. I love Andrew Ng.
@@99dynasty i will buy it
@@99dynasty 1
@Potato Potato was gonna type the same thing haha
yes have heard those before every time a topic came up that Andrew feared might be too much for his students.
recall his lectures
I teared up when Andrew said “someone allow me the good fortune of helping them a little bit on their path to their dreams”. It restores my faith in humanity when I hear influential people like him with this much humility.
I took Ng's class on Coursera years ago. It switched my career trajectory from academic to industry, and spurred my interest in making things. Awesome to see this interview!
Robert Lewis I’m late to the show and going thru it now but yes. Changing my career direction now
Hi, I take the same class. Can you recommend some next course or anything else?
@@unreactive I really enjoyed reading DataSmart by John Foreman. I also recommend the website "fast.ai" (I found out from Lex). Finally, there are good courses on specific techniques, programming languages, etc, on how to use ML on udemy.com. Udemy seems more practical, and has taught me how to actually implement some things and deploy them for use in the cloud.
Awesome, thanks!
How much is it?
One of the people that inspired me to do artificial intelligence in areas that matter. Without Coursera I would have not gotten where I am today. Thank you for this conversation Lex!
Splendid!
did you finish it? I'm at course 4 week 4 in deeplearning.ai
Is it really that good?
Im a physician from Brazil learning ai and this fine man helped me a lot thank you sir I salute you
You’re making these vids almost as fast as I can watch them! Thanks so much my favorite podcast and my favorite thing on youtube.
I currently work as a computer vision engineer. I got hired having only a BS in Math and zero ML experience. My first assignment was to take Andrew's ML/DL classes on Coursera. Invaluable. Thank you!!
Would you still recommend it for beginners today?
This man got a great voiceover career available at anytime.
53:30 Andrew's Recommendation on how to learn more deeply with handwritten notes
Handwritten notes is better than typing but I thought that is not surprising. That being said, taking notes In itself is not that good for revision/recall. Active recall and spaced repetition are the best for that
I have to agree with this, even while learning programming, writing things down actually helps a lot.
I found this basic technique very useful for studying kanji while learning Japanese. My theory at the time was that the extra effort I had to expend was forcing my brain to focus more on the content rather than going into autopilot mode. However, I also found that intervalled repetition was a big part in the learning process.
Agreed. Active recall (trying to answer own questions with the solution covered) and spaced repetitions are best. But trying to summarize a point in your own words in almost as good. I think Andrew was pointing to that aspect.
27:34 "If you want to make a breakthrough, you sometimes have to do something before it becomes popular". What a humble man :)
He's the best teacher I have met so far. He teaches all the concepts to its very core and many times even predicts his student's doubts and clears it(really felt as if i was into his class where all my doubts were getting cleared) . His teaching made me realize my passion for AI, ML which decided the trajectory of my career. He teaches such complex concepts in such a simplified way (that's the beauty of teacher).At one point of time I could literally visualize statistics making my foundations clear(especially gradient descent) . It will truly be my fortune and blessing to ever meet my mentor in person. Thanks Lex Fridman for this video where we all could know more about Prof
Andrew Ng.
Man I love your podcasts..Andrew Ng is such a great educator. His deep learning specialization on coursera is what got me into machine learning. His Stanford machine learning lecture playlist on RUclips is awesome as well.
This is so awesome! Thanks for having Andrew on.
I've been waiting for this!!! Thanks bro!
This one I will listen full!
I was long waiting for you to bring him!!
Thank you so much for making this available Lex.
Andrew literally changed the direction of my life with his Stanford ML course and subsequent work with Coursera (which happened to introduce me to Geoffrey Hinton's deep learning course).
His work ethic and drive to help the most people has helped reshape the face of education and I don't doubt it will leave a lasting positive impact on this world.
Great, enjoying my dinner at the moment. time to feed my brain some information also! Let's go.
Fantastic interview. I took Andrew's ML course and am still deeply impressed by his pedagogical abilities. He is very observant and masters the art of passing on his knowledge. What an enrichtment these new distance learning WBT methods are for society! Perhaps the biggest improvement in this still new century for quality of life and fairness regarding access to knowledge and skills. Thank you from a humble admirer.
This is a very inspirational interview. Thanks for it, Lex!
thanks for releasing this have been waiting for quiet sometime
I have taken several of Andrew Ng's courses. I have thoroughly enjoyed them and learned so much. He is a great teacher and a very likeable person.
As a GenX'er I enjoyed this interview so much. It made me actually feel like I was experiencing the internet we were promised decades ago. A well thought out and intelligent conversation that, in the end, leaves the viewer in a state better than when they first pressed the play button. Well done.
So inspiring! Thanks a lot for this and all the other interviews!
Listening to Andrew Ng is so nostalgic lol I remember how much I enjoyed my first step into AI with his ML course
He explains very eloquently actually! Thanks for the interview, Lex!
Thanks for being considerate in the way you do ads. Much appreciated.
Lex, thank you for educating the world for free, I am grateful beyond belief.
Fantastic! What I like about your casts is the choice of people to talk to and meaningful line of discussion. They elevate our understanding a little bit, looking forward to every episode!!
Andrew Ng is my Guru.. This is first podcast I listed fully in a silent room
Thank you so much for making this available Lex.
Both are wearing their iconic dresses! XD
Those are their super hero suits
@@BasharAlkhalili brilliant people seem creatures of habbit. what to wear is not just a distraction, it's irrelevent to them (i'm thinking of einstein in those effeminate sandals)
This man has pulled some strings of my soul. And it's really motivate me for dive in this field. Thank you, Lex
Loved this video, I found it really interesting to hear Andrew Ng draw the analogy of litteracy to data science or coding skills.
This is one of the most impactful podcasts I've heard! Thanks so much, Lex & Andrew!
Единственный подкаст, который я могу слушать постоянно. Лекс, спасибо за первоклассный контент! В сети вряд ли найдется другой такой же впечатляющий источник философского анализа темы искусственного интеллекта. Как говорится, instant classic )
no words can describe how much i love andrew
Thank Lex for interesting conversation !
Thank Andrew, I'm one of your million students !
Fantastic, just in time for my 2hr drive
Nice listening/watching to both of you. Thank you.
Hands down to Andrew. Great mind with such a down to earth personality. Just love composed and calm vibe that he gives out when he talks as well.
Andrew Ng is a great teacher, great person, great(to the power of n) to me and my friends. I request this great human being to build systems in healthcare and life sciences that would transform the lives of people around the world, particularly the poor-needy and under privileged.
Thanks, Lex for this amazing podcast. Thanks Prof. Ng for thoughtful advices. What I liked best - Try to study 2 Research Papers every week, make a habit of something. Think about who would be your peers before joining a team. Thanks.
I am watching it 3 years later and I must say this is a great interview. Great work Lex and Andrew.
So exciting to see this talk. Two years ago, I started learning ML, the first series of courses about ML I watched were made by NG. He is my first ML/DL influence. He is such a nice educator.
Finally ! Andrew Ng.
Thanks Lex.
The conversation is quite informative.
Andrew NG is the one who introduced me to the world of Machine learning through Coursera few years back. I think it takes a special kind of expertise to make complex things sound simple. Andrew NG truly has done a great job at explaining basics of ML to all of us. He truly is the father of Machine/Deep learning.
Andrew Ngs point about reframing the problem to solve it yourself was extremely insightful. I'm going to try this with my nephews when we make puzzles - I'll show him initially how some things can connect (not all) and have him "reframe it" to solve it himself! - super excited
Thank you Andrew Ng, you are the best. I'm learning so much from you.
amazing season...i am learning a lot in this podcast ..thank you so much
Thank you Lex Fridman for this great podcast!
Lex, More love to you bud. Enrolled in Coursera and signed up for Andrew's course because of this!!! THANKS!
Great job. I have been totally inspired by Andrew Ng two years ago.
A podcast... waited for quite long time. Thanks Lex ✌️
Great man. Really enjoyed every course taught by him.
His dedication reaches thousands or even millions of people across the world.
Andrew NG has lot of fans in INDIA 😊
Two beasts in a conversation. Awesome stuff!
It's an amazing time to be alive. I'm fascinated by Auudia's new Vodacast. Perfect podcast compliment.
Thanks for having Andrew on. He is above all else very authentic in his journey.
💚🤍love this conversation, the warmth and gentleness from the brilliant scientists, thanks
Brilliant conversation. Thanks to you both.
Andrew Ng and Lex Fridaman voices are so soothing
Man this is one of the best guests that we can have in the field !!! 😂😂 that was awesome that you could get this generous man!!!! Such a wonderful human you to put him on the mic and him from providing such words
Yes, was waiting for this one!
A special interview of all. Thanks lex.
Hellz yeah! Big Ng on the podcast!
Amazing to see two great minds talk! Especially Andrew Ng who introduced me to the field of machine learning!
I have been waiting for this long ago. Thanks Andrew. Thanks Lex.
Great interview! Andrew is a true legend. My only complaint is that I didn't hear him say "Concretely" once!
so much love and respect for this guy!
Thank you for this! I'm a beginner and love this type of knowledge and insight!
Fascinating interview!
Thank you for Lex for this amazing podcast ! ! !
Great interview. Thanks Andrew & Lex!
I was waiting for this Podcast. Guruji!
Wow. Another amazing guest.
01:01:00 That's a huge piece of advice! And I couldn't agree more. People tend to get attracted to the logo - but the experience individuals have is so broad and ranges from poor to incredible even inside Google, Microsoft, Facebook and all of those famous tech companies.
I really enjoyed this interview done by you and Andrew.
Tremendous !!
Thanks heaps Lex !
Love the changes Lex. Thanks for taking suggestions man. ^_^ Keep doing you!
Good to see so many of the listeners took up ML/DL courses from Andrew Ng. I am one of them and actually been doing professional work for more than a year. Without Andrew Ng and Coursera, I would have never gotten into ML.
I came from no formal education on computer science or AI, took Ng's machine learning MOOC, and fast forward a few years and now I'm in my 2nd year of my CS PhD with a focus on AI.
Wait, so you got into CS PhD without doing the CS undergrad?
I had an undergraduate degree in physics and master's in mechanical engineering@@seetsamolapo5600
This is just fantastic. :-D Love both guys, they are great
Thank you Andrew Ng and Lex Fridman
such a great mentor, he makes that's everything is so easy.
Andrew's videos on RUclips are goated. 🐐
This week has been awesome! First Scott Aaronson and now Andrew Ng :).
Lex, it'd be really cool if you could get someone working on the theoretical aspects of Machine Learning, like Sanjeev Arora, Ankur Moitra or Costis Daskalakis... It'd be super interesting to see the conversation between experts on two different sides of the same coin; applied and theoretical.
Andrew is one of those people whos impact on the world is impossible to estimate. He's a living legend. Born to be an educator.
Thanx for bringing Andrew
I've become one of Andrew's fans, especially after starting Deep Learning specialization and finishing the first course. It has changed my life.
I think this might have changed my life!!! Thank you!!
Wow!!! Thank you so much!!!
I'm used to watching his videos at 1.5x... so hearing his voice at normal speed tripped me out lol. Joking aside, what a great interview with a true pioneer and legend. Thanks so much Lex!
I love the part where they talk about handwritten notes over typed notes. The fact that you have to "encode" the information in order to write it down in fewer words [rather than be able to transcribe verbatim when typing] is key. When we encode the data for our notes, we are learning a set of nonlinear transformations which connect our encoded notes and the original, larger resolution, message. In performing such a task, we learn features about the message, thus gaining what we call "a deeper understanding." The brain is a wonderful machine.
I love the guy! He influenced my knowledge and path, as well as many of you, for the better!
Thank You So Much. Keep up the good work.
Since you talked about reinforcement learning, what type of path would you think will take us to AGI?
Is it a system of integrated solutions from different theories and aspects? Or is it a single learning framework, for example deep reinforcement learning?