Daphne Koller: Biomedicine and Machine Learning | Lex Fridman Podcast #93
HTML-код
- Опубликовано: 6 июн 2024
- Daphne Koller is a professor of computer science at Stanford University, a co-founder of Coursera with Andrew Ng and Founder and CEO of insitro, a company at the intersection of machine learning and biomedicine.
Support this podcast by signing up with these sponsors:
- Cash App - use code "LexPodcast" and download:
- Cash App (App Store): apple.co/2sPrUHe
- Cash App (Google Play): bit.ly/2MlvP5w
EPISODE LINKS:
Daphne's Twitter: / daphnekoller
Daphne's Website: ai.stanford.edu/users/koller/...
Insitro: insitro.com
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
OUTLINE:
0:00 - Introduction
2:22 - Will we one day cure all disease?
6:31 - Longevity
10:16 - Role of machine learning in treating diseases
13:05 - A personal journey to medicine
16:25 - Insitro and disease-in-a-dish models
33:25 - What diseases can be helped with disease-in-a-dish approaches?
36:43 - Coursera and education
49:04 - Advice to people interested in AI
50:52 - Beautiful idea in deep learning
55:10 - Uncertainty in AI
58:29 - AGI and AI safety
1:06:52 - Are most people good?
1:09:04 - Meaning of life
CONNECT:
- Subscribe to this RUclips channel
- Twitter: / lexfridman
- LinkedIn: / lexfridman
- Facebook: / lexfridmanpage
- Instagram: / lexfridman
- Medium: / lexfridman
- Support on Patreon: / lexfridman 
Наука
![Dimension 20: Never Stop Blowing Up Trailer [Dropout Exclusive Series]](http://i.ytimg.com/vi/xYeXpbHnGEM/mqdefault.jpg)
I really enjoyed this conversation with Daphne. Here's the outline:
0:00 - Introduction
2:22 - Will we one day cure all disease?
6:31 - Longevity
10:16 - Role of machine learning in treating diseases
13:05 - A personal journey to medicine
16:25 - Insitro and disease-in-a-dish models
33:25 - What diseases can be helped with disease-in-a-dish approaches?
36:43 - Coursera and education
49:04 - Advice to people interested in AI
50:52 - Beautiful idea in deep learning
55:10 - Uncertainty in AI
58:29 - AGI and AI safety
1:06:52 - Are most people good?
1:09:04 - Meaning of life
Lex Fridman beautiful contribution!
Awesome! Thanks Lex
These are so useful Lex thank you
excellent
@Jeff finally someone who was wondering the same as me .come on lex please tell us what watch it is you wear I'm a little bit of a collector myself and I also make ww2 style fliger watchs with hand wind 6497 movement lumed hands and dial or applied index and no lumed options lex I would love to make you one email me 2e0njk@googlemail.com
This Professor is incredible. I could listen to her all day. Extremely clear and to the point. Thanks Lex!
You beat me to it. I was going to post the same. No extra words and perfectly clear. Would love to hear more of her podcasts or lectures.
I was dreaming about this conversation! My favorite scientist and entrepreneur, Daphne Koller, talks with my favorite podcast author, Lex Fridman! This is just awesome. I cancel all of work until I finish watching it.
Can’t get enough of all these podcasts, listened to so many of them several times, thanks for everything Lex!
This is one of those interviews that make you feel hopeful, inspired, and motivated. Thank you.
There were so many ideas I liked, but they can be summed up with Daphne's comment at 71:00: "use your life to make the world a better place." ❤️Beautiful.
People do make the world a better place sadly just for themselves.
@@arunhazarayou1 Even the most selfless acts can ultimately be seen as selfish if we analyze them long enough. For example, if a person gives up their life to protect someone they love, we can say they did it to spare themselves the pain of living without their loved one. Or if that person gives up their life to protect a complete stranger, we can say they did it to spare themselves the pain of living with a guilty conscience.
It would be a pity though to say that people are selfish when trying to make the world a better place. First, the personal gain for some of these people is almost non-existent. And second, I actually have great respect for anyone who does some act of kindness, even if it benefited them. They could have done nothing at all. Instead, they chose to contribute.
Like Emily Dickinson said, “If I can stop one heart from breaking, I shall not live in vain.” Helping even one person is an admirable act. ❤️
@@BiancaAguglia
I understand that there are almost no selfless acts but still there are degrees of selfishness. The way your amazing and kind brain puts up effort in dissecting almost heroic goodness you know in the same universe there are minds putting same kind of energy making their acts more and more selfish. Eg. all the poverty in this world and hoarding of wealth on the other side.
I guess it's people like you who keeps our world from going all dark. Or some very capable group to oppress all else I think it's humans like you.
Thank you for your efforts. Couldn't belive for a moment that you were real the way you think.
I have a hard time balancing selfishness, found one needs atleast some to survive.
One time I felt so much for the bacterial colony dying on my petri dish eating each other but couldn't do much.
It is an interesting statement for her to make. She is a billionaire. I think its strange that nobody else has mentioned this. I am not saying that is a negative thing or inherently makes he "bad" but I would have liked that to be mentioned in a sense. I dont know how accurate of a view you can have on whats "good" for the world if you have ridiculously vast wealth.
I like it that you ask your guests about the meaning of life. Very important question. Always was. Always is.
The answer is simple:
you were never _always was_
you will never be _always is_
you are just the bit in between
that has a short was
an instant of life now
and a short will be
but no more.
It seems the meaning is often the same. Its the burden and struggles that give meaning and happiness. Standing in the house of order you built from bricks of chaos.
My mind is ready 😬🤯😁🔥 The intersection of Machine Learning and Medicine is one of the big reason I was inspired to dive into the world of Computer Science and Engineering. The synergy here has the potential for enormous impact! Thank you, Lex!😁🙏🏽
I've spent hours on her lectures on Probabilistic Graphical Models.
She's a great teacher too! 😄
Link please ?
She is terrific. Thanks for another fantastic episode, Lex.
Makes me wishing I was in America and able to pursue research opportunities in these areas.
You are _not_ alone.
Daphne Koller is the best! Her book on probabilistic graphical models is fantastic!
But it's not Daphne Koller _at_ her best, because that would simply scare people. I used to follow her work from the 1990s a bit, which I discovered around the time of her MacArthur Fellowship. Serious, substantive work. Masters degree at 18 and doesn't seem to have taken a day off, since.
2010: Newsweek's 10 Most Important People
2013: Time magazine's 100 Most Influential People
2014: Fast Company's Most Creative People in Business
There _was_ another woman at Stanford in roughly the same cognitive zip code as Daphne: Maryam Mirzakhani, an Iranian mathematician who won a Fields Medal, but then tragically passed away from breast cancer at age 40.
Great job, Lex. This podcast is quickly becoming one of the best!
I was dreaming that you'll get to interview her! Thank you from the bottom bottom of my hurt. I'm watching all your casts, Lex. Please, keep going. you're the best in the podcast space.
The quality of these conversations and the fact you keep them very orderly is incredibly valuable to me. You're doing a great service to the world with your work Lex. Thank you.
This was an amazing interview. I'm grateful and really it's just awesome to see how far Lex has come as an interviewer.
Incredible that you get this type of informational quality for free. Thank you Lex!
Coincidently, I'm taking a Coursea course and absolutely loving the class. Knowing the background has made it even more profound, thank you, Daphne, for sharing and your efforts to fine-tune the classes (time and efficiency). Online class opportunities have opened doors that as you mentioned, "at this life juncture" brick and mortar classroom opportunities seem limiting. I also love your perspective on diversifying STEM fields. Lex your guest speakers definitely expose your viewers with timely topics and a wide-demographic of amazingly genuine intellects to expand their minds. Thank you!
These podcasts are just what I need during lockdown. Thanks
This was delightful! Thank you.
Amazing conversation! Great podcast.
Really felt deeply thinking on how well we can connect the planet with all kinds of problems and challenges through machine learning.
Thanks for the great video. It was really inspiring and aspiring to hear most tremendous people thoughts and works towards the benefit of mankind.
Thank you Lex - I hope this inspires many people in the life-sciences to adopt machine learning in their research including myself.
Brilliant interview. Coursera had a huge impact on my life and I have never heard about the co-founder Daphne Koller. The deep learning courses from Andrew Ng are without a doubt the best material for understanding the concepts and maths behind deep neural networks. I have to take Daphnes course on Probabilistic Graphical Models now.
I’ve been listening to your podcasts for so long, while I work. That I can tel when the podcast is almost over when you ask your guests what the meaning of life is
Educators like Daphne gives me hope for the future.
Amazing conversation,thank you
Dear Mr Fridman and Ms Koller, Thank you Mr Fridman for a terrific interview, and thank you Ms Koller for what Mr Fridman said at the end about it being an honor to hear about your life and work.
Great podcast, thanks for share with us the different ways to see the potentiality of Deep Learning 👍👍👏
I really love the quotes you read at the end, great content as always
The talk I was waiting for. Cheers again Lex!
I've been waiting for this podcast :-)
1:06:52 - Not everyone emphasizes such an important point with so much sincerity. She's great!!
Good job Lex, nailed it as usual!
This lady is one of the most important persons in AI and Biology ! Lex please do another round !
Great job on the podcast as usual. hey once this COVID business is done try to get Robert Sapolsky on the podcast.
Incredible! Thanks Lex.
Impressed by how Stanford's commitment to improving educational quality was a contributing factor to Coursera's success.
Such a cool guest! She is incredible!
The applications of AI/ML to healthcare and medicine is a fascinating topic! The future potential of this area is very exiting! Great conversation!
very inspirational work and wisdom. Ty
Love seeing brilliant women in your podcast! :) Truly inspiring.
Wow, imagine being her student, or working with her.
She is so passionate about her work.
Top notch content! Love to hear from researchers on the front edge. I am part of a FMT facebook group, the stories of autoimmune diseases, even autism in children being cured by altering the gut bacteria. People so desperate that they are experimenting on themselves! That where she should be pointing her very powerful ML engine! That data will be expensive to obtain, until Hight Throughput DNA sequencing drops in price. Being Canadian, I would like to see you interview Jeff Hinton, the whole voice in the wilderness story is satisfying! Not to mention, I like your questions!
Beautiful. That ending made me cry.
Awesome podcast 👍
I'm taking Dr.Koller's course on Probabalistic Graphical Models right now. Really enjoying it so far :).
Fantastic as always dude 👌
She cranks out these really long sentences and uses these uncommon words in sometimes uncommon ways, yet with no background in machine learning or biology, I am able to follow the meaninig! That's always a sign that the speaker undersdands the field really well. What a pleasent (so far) 30 minutes of horizon-expanding information.
Give it welly, Jazzo.
@@tonyduncan9852 Never heard that english expression before. But I did give it welly, didn't I?
@@jazzochannel Indubitably. :)
Thank for this video. Important stuff to cover!
She is such a pearl of our world
Yet another awesome podcast, thanks Lex! If you're able to see this comment, could it be possible that you would gather a book list on your website of interesting books on machine learning, brain, conciousness... The stuff you read in general?
Ive watched most of Rogan and Petersons content available on RUclips. They boosted my confidence and curiosity for some intellectual conversations.. yours is awesome but will quickly let one know exactly how much is lacking in current education. Some might as well be in Sumerian. From what my caveman brain and HS education can interpret your shows really are great. Thank you
one of the few people actually use ml rationally in biomedical ! deep understanding of both math and biology is crucial!
Great guest and great conversation!
Sad that this is one of the least amount of views on your channel. Need a round #2, now with a bigger subscriber base you have built. This is an amazing thinker tackling very important problems!
i wonder how much ammo Lex has put in his "In case of COVID, release these" box
I already recorded a conversation with everyone who has ever lived and wrote a program in Python to occasionally publish new episodes. I am no longer alive and what you think of as Lex Fridman is an illusion: a Python script automatically publishing podcasts indefinitely until the heat death of the Universe.
@@lexfridman And who said Russians don't have a sense of humor!
Lex Fridman 😂😂
@@Shoveltime198 Whoever said that Russians don't have a sense of humor is no longer with us. Rest in peace.
@@lexfridman I guess you also created a NN model making deep fake interviews with people who don't even exist
She is what Elizabeth Holmes pretended to be, a genius heading up a silicon valley biomed startup.
She was a total sociopath! She changed her corporate schedule to match Steve Jobs. She was Jobs, without the talent!
Robert Callaghan yeah man , that deep voice tho 🤦♀️
@@bumberClart1000 She was caught lowering it.
@@bumberClart1000 Bumped into a cool series, tinyML tied to an Oreilly book. and video series, alot like sendtex ruclips.net/video/8N6-WQsxwxA/видео.html
@@7Trident3 Jobs wasn't particularly talented. All of his success comes from selling another man's product and then prestige pricing an mp3 player. Somehow the masses then turned him into a cult icon.
Спасибо за такой замечательный контент!
A brilliant, strong, educated, & independent woman. A role model for teens who aspire to change our Earth with education and technologies.
Yoo Lexingrad, I think you'd love talking to David Sinclair considering your questions here.
She's freaking awesome btw. Working on both online education and bioinformatics which are probably gonna be hugely disruptive in the upcoming years.
Id love to see some of your projects lex, tensorflow?
Amazing guest. Her book on PGMs is great!
Lex It is early morning hear in India. Its heavy raining here. And I am watching a video which I like too much.
i watched the whole alex garland podcast, great convo.
Bet she wasn’t surprised by covid-19! Thanks for airings this ,it should have been aired earlier ,better late than never !🙏
Could you please only upload full episodes?
I didn't mind that this stayed on fairly general terrain. And I really enjoyed the way Daphne is relentlessly articulate. But then you've got Lex asking the questions, and there were moments where he was qualified to ask an especially important question, yet he pulled the chute. The big one I noticed was when Daphne was talking about the lack of robustness in our present ML systems, commenting to the effect that "small tweaks can completely change the outputs". But this in some sense is an automatic consequence of working in high-dimensional spaces: there are always going to be inputs on the knife-edge. You simply can't avoid having those interior manifolds. However, if those knife-edge manifolds are forced to be relatively high-dimensional objects, they sort of disappear as measurable sets. But is it really feasibly to pursue a form or program of robustness where we have some kind of guarantee that the vast majority of knife-edge interior manifolds amount to nigh-immeasurable sets? I know I'm not quite using insider language here, but that's why I'm _listening_ to this podcast, rather than hosting it myself. It really wouldn't have hurt things to let Daphne field at least one hardball, because I'm pretty sure she's up to it.
It’s like all the awesome ppl in the world are in the comment section ..... you guys make me hopeful for humans.... stay awesome fellas.
I just finished a course on Coursera .. opened YT and this was the first video on it 😂
What a wonderful woman.
Imagine if we reach the stage of machine learning and AI when we can do the equivalent of 10 years or more of human research in a day.
As I see it, the overall goal course seems to be for people to live in good health and shape for as long as they wish to live.
Lex can we have a round two please ?
Her answer on AGI was excellent
Wonderful
Great!
1:07:58 ALL want the perception of " right doings/actions" even the MOST DECEPTIVE
What a beautiful and inspiring human being!
10:02, just hear her call 150, she really loves it.
She's amazing!
Finally they're keeping track of patients data at hospitals with a single folder ascribed to them on a computer... this will aide science and society greatly.. why did it take so long😥
(Sigh) Sadly people are more interested in killing each other than curing each other
Years down the line, this podcast will be seen as seminal in being able to persuade narrow natural intelligence experts to reflect on natural general intelligence as it relates to the nature of reality, consciousness and the meaning of life.
Not sure if any other platform has been able to do this, but I'd bet that a meta-level appreciation of NGI is critical to the development of AGI.
In the future u will get to choose which ai robot assistant is right for u amazon, apple, google, fb or microsoft or u can get the Chinese version Alibaba, Tencent and Baidu
Amazing person.
AI/AGI community love!! 🤖🧡
from caption to video was like time travel
59:43 boy let me tease apart different parts of that question... I'm like that reminds me of my tutor in schl
One of the more interesting presentation I've listened too. One thing sir, isnt it more accurate to say the professor is in the treatment business NOT the curing of any diseases. Otherwise great show
🕶🎥
Id like to learn more from these types of things. We get "we're nowhere near solving this problem" but not a description of the obstacles and what makes them difficult. I also think Lex is a skilled enough interviewer to pin people on evasive answers without being a kathy newman.
super cool
hero
6:52 thanks God I am turning 40 this year
I wish you a long life sir.
“In modern medicine, Dr. Atul Gawande said, there are 13,600 diagnoses, or ways in which the human body can fail, and no patient comes in with just one diagnosis at a time. Now more than 6,000 drugs can be prescribed, and 4,000 medical and surgical procedures can be performed.”
Such a shame this interview happened before the pandemic took off, the topics of online learning and health have changed a lot in the past year. I hope she can come back on again to give a new perspective
Both, Lex! BOTH!!!:)
If you can get Dr. fei fei li on your podcast that’ll be superb! She’s a hero to what she contributed to computer vision!
Very nice interview!
Of course deseases are part of the world and probably on the long scale people need / want to experience certain body states to become old and experience the variety of life. Because in later states / ages deseases have a more fatal impact. This could be done in controled phases and shows how important it is to learn more about it . But of course why complexity hold this kind of valleys? (Deseases? Wars? Animal testing? Strange behaviours?)
If you devide the good and the bad answers into use now and use later cases could solve some confusion.
"Machines learning now and soon I will be born to ask you. Who you want to be?"
*Smith is a beautiful programm, but you can loose control if everybody is just doing wat he wants to and not what is neccessary. Because neural networks are strong and maybe machines need this to navigate the universe. :) And of course there are always people hiking for the truth!
We have not progressed past a world where people can't do really terrible things and still be thought well of
1:11:11 It's not pointed out nearly often enough that to leave the world a better place you have to outperform your initial conditions (which Koller gestures toward using the proxy "privilege") and you can't be alone in this. Progress only occurs when the _majority_ of people in society outperform their initial conditions. Moreover, it doesn't take much defection from people who begin with enormous privilege to undo almost all the hard-won progress single-handedly. (Not that there's anything wrong with that, if you're on Team Red.)
This must be the most widely travelled MIC + MIC stand + XLR Cable in the world ... :)
i wanna live as long as the stars do
Nice interview, I have worked in the field of AI and medicine in the past as a researcher. (Some nonsense remark from my side, just a quick association: She sounds a bit like "Agent Smith" from the Matrix... She's perhaps his "good" female counterpart? ;-) )