吴恩达 机器学习课程 37 吴恩达与李飞飞谈以人为本的AI #机器学习 #基础教程

[Music]

Hi, I'm delighted to have with us here

today my old friend professor FE Lee. Fe

is a professor of computer science at

Stanford University and also co-director

of HAI the human- centered AI institute

and previously she also was responsible

for AI at Google cloud as a chief

scientist for the division. It's great

to have you here.

>> Thank you, Andrew. Very happy to be

here.

>> So, I guess actually, how long have we

known each other? I've lost track.

>> Definitely more than a decade. I mean,

I've known your work, right, before we

even met. And uh I came to Stanford

2009, but we started talking 2007, so

>> 15 years. And I actually still have very

clear memories of um how stressful it

was when collectively, you know, bunch

of us, me, Chris Manning, a bunch of us

were trying to figure out how to recruit

you to come to Stanford.

>> It wasn't hard. I just needed to sort

out my students and life, but uh it's

hard to resist Stanford.

>> Wasn't it really great having you as a

friend and colleague here for

>> Yeah, me too. It's been a long time and

uh we're very lucky to be the generation

seeing AI's great progress.

>> Hey, so there was something about your

background I always found inspiring

which is you know today people are

entering AI from all walks of life and

um sometimes people still wonder oh I

majored in something or other is AI a

right path for me. So I thought one of

the most interesting parts of your

background was that you actually started

out not studying computer science or AI

but you started out studying physics and

then had this path to becoming you know

one of the most globally recognizable AI

scientists. But so how did that how did

you make that switch from physics to AI?

>> Right. Well that's a great question

Andrew especially both of us are

passionate about young people's uh

future and their uh um they come into

the world of AI. The truth is if I could

enter AI back then more than 20 years

ago, today anybody can enter AI because

AI has become such a um prevalent and

and globally impactful technology. But

myself um maybe I was an accident. So um

I have always been um a physics kid or

STEM STEM kids. Um I'm sure you were

too. But physics was my passion all the

way through you know um middle school,

high school, college. So I went to

Princeton and majored in physics. And uh

one thing physics has taught me till

today is really the passion for asking

big questions, the the passion for

seeking north stars and uh I was really

having fun as a physics student at uh

Princeton. And one thing I did was

reading up um uh stories and just

writings of great physicists of the 20th

century. I just hear about what they

think about, you know, the world,

especially people like Albert Einstein,

Roger Penrose, um, you know, Erin

Schroinger. And it was really funny to

notice that many of the writings towards

the later half of the career of these

great physicists were not about just the

atomic world or the the physical world

but ponderings about equally audacious

questions like life intelligence like

human conditions. you know, Schroinger

wrote this book, what is life? And uh

Roger Penrose wrote this book, Emperor's

New um um u new mind, right? And uh that

really got me very curious about uh the

topic of intelligence. So I one thing

led to another during college time. I

did intern at a couple of neuroscience

uh labs and and especially vision

related. And I was like, "Wow, this is

just as audacious a question to ask the

beginning of the universe or what is

matter made of?" And that uh got me to

switch from undergraduate degree in

physics to graduate degree in uh AI.

Even though I don't know about you,

during our time um AI was a dirty word.

>> It was AI winter. So it was more machine

learning and computer vision than and

computational neuroscience.

>> Yeah, I know. Honestly, I think when

when I was in undergrad, I was too busy

writing code. I just, you know, managed

to blindly ignore the AI winter and just

kept on coding.

>> Yeah. Well, I was too busy solving PD

equations.

>> Hey, and so actually, do you do you have

an audacious question now?

>> Yes, my audacious question is still

intelligence. I think since Alan Turing,

humanity has not fully understand what

is the fundamental computing principles

behind intelligence.

I you know we um we we today we use the

words AI, we use the word AGI but at the

end of the day I still dream of a set of

simple equations or simple principles

that can define the the process of

intelligence whether it's animal

intelligence or machine intelligence and

this is similar to physics for example

many people have drawn the analogy of

flying right are we replicating birds

flying or are we building airplane? And

a lot of people ask the question of the

relationship between AI and brain and to

me whether we're building um a bird or

replicating a bird or building an

airplane at the end of the day

aerodynamics and physics that govern the

the the the process of flying and I do

believe one day we'll discover that.

Yeah, I sometimes think about this, you

know, one learning algorithm hypothesis.

Could a lot of intelligence, maybe not

all, but a lot of it be explained by one

or a very simple machine learning

principles. And it feels like we're

still so far from cracking that nut. But

in the weekends when I have spare time

when I think about learning algorithms

and where they could go, this is one of

the things I still, you know, am excited

about, right? Just thinking about

>> I totally agree. I still feel like we

are pre-Newtonian

if we're doing physics analogy. Uh

before Newton, there has been great

physics, great physicist, a lot of

phenomenology, a lot of studies of how

the the the the astral bodies move and

all that. But it was Newton who start to

write the very simple laws. And I think

we are still going through that very

exciting coming of age of AI is a basic

science and uh we're we're pre-new

Newton. Um in my opinion

>> it's really nice to hear you talk about

how despite machine learning and AI

having come so far it still feels like

there are a lot more unanswered

questions a lot more work to be done by

maybe some of the people joining the

field today than work that's already

been done.

>> Absolutely. Absolutely. I mean let's

let's calculate it's only what 60 years

about. It's a very naent field modern

his physics and chemistry and biology

are all hundreds of years right so I

think uh it is very um it is very

exciting to be entering the the field of

intu uh science of intelligence and and

u and studying AI today. Yeah, actually

that's good. I think I remember chatting

with the late um professor John McCarthy

who had coined the term artificial

intelligence

>> and boy the field has changed since when

you know he conceived of it at the at

the workshop and came up the term AI but

maybe another 10 years from now you know

maybe someone watching this will come

with a new set of ideas and then

>> absolutely

>> we'll be saying boy AI sure is different

than what you know you and I thought it

would be that that's an exciting future

to build towards

>> yeah I'm sure Newton would have not

dreamed of Einstein. So, you know, our

our evolution of science uh sometimes

takes strides, sometimes takes a while,

and I think we're absolutely in a in an

exciting phase of AI right now.

>> You know, it's interesting hearing you

paint this grand vision for AI. Going

back a little bit, there was one other

piece of your background that I found,

you know, inspiring, which is um when

you're just getting started, I've heard

you speak about how, you know, you're a

physics student, but not only that, you

also you're also running a laundromat to

pay for school. And so, just tell tell

me more about tell tell us more about

that.

>> Well, yeah. Um so I I came to this

country to America to New Jersey

actually when I was 15 and uh one thing

great about being in New Jersey is it

was close to uh Princeton. So I I often

just take a weekend trip with my parents

and to admire the the place where

Einstein spent most of his career in the

latter half of his life. But uh you know

with typical immigrant life and uh it

was tough and uh by the time I entered

Princeton my parents didn't speak

English um and uh one thing led to

another. It turns out uh running a dry

cleaner might be the best option for my

family especially for me to lead that

business because it's a weekend

business. If it's a weekday business, it

would be hard for me to be a student.

And uh um it's actually, believe it or

not, running a dry cleaning shop is very

machine heavy, which is good for a STEM

STEM student like me. Um so we uh

decided to open a dry cleaner shop in a

small town in New Jersey called Pipony,

New Jersey. It turned out we were

physically not too far from Bell Labs

and where lots of uh early convolutional

neuronet network uh uh research was

happening but I had no idea

>> actually a summer intern at AT&T Bell

way back

>> with Rob Shaper

>> um with Michael Karns was my mentor and

Rob Shaperi inventor boosting great

algorithms.

>> See you're coding AI. I was trying to

cling to no that was only much later

only only much later in my life did I

did I start interning.

>> Yeah. Um and then uh um it was seven

years I I did that for uh the entire

undergrad and most of my grad school and

I hire my parents

>> and yeah.

>> Yeah. No, that's really inspiring. I

think it, you know, I I I know you've

been brilliant at doing exciting work

all your life and I think the the story

of, you know, running a laundromat to

globally prominent computer scientists.

I hope that that that inspires some

people watching this that no matter

where you are, there's plenty of room

for everyone.

>> Yeah. Don't even know this. Um, my high

school job was a office admin. Um, and

so to this day, I remember doing a lot

of photocopying and and and the exciting

part was using the shredder. That was

the glamorous part, but I was doing so

much photo copying in high school. I

thought, boy, if only I could build a

robot to do this fellow coffee, maybe I

could do something else.

>> Did you succeed?

>> Oh, no. Still working on it. I don't

know. We'll see. We'll see. Yeah. And

then um you know when people think about

you and your the work you've done one of

the huge successes everyone thinks about

is image net right where um hub

established early benchmark for computer

vision. It was really completely

instrumental to the modern rise of deep

learning and computer vision. Um one

thing I bet not many people know about

is how you actually got started on

imageet. So you tell tell us the origin

story of of of imageet. Yeah. Well,

Andrew, that's a that's a good question

because a lot of people see ImageNet as

just labeling a ton of images, but where

we began was really going after a

Northstar brings back my physics

background. So, when I enter grad

school, when did you enter grad school?

Which year?

>> Uh 97.

>> Okay. I was three years later than you.

2000. And that was a very exciting

period because I was in the computer

vision and computational neuroscience

lab of Petro Perona and Kristoff K at

Caltech and leading up to that there has

been first of all two things was very

exciting. one is that the world of uh AI

at this po at that point wasn't called

AI computer vision or natural language

processing has found its um lingua defco

it's uh uh machine learning statistical

modeling as a new tool has emerged right

I mean it's been around

>> and and I remember when the idea of

applying machine learning to computer

vision that was like a controversial

thing

>> right and I was the first generation of

graduate students who were embraing

releasing all the base net all the

inference algorithms and and all that

and uh and that was one exciting

happening. A second exciting happening

that most people don't know and don't

appreciate is the u couple of decades

probably more than two two or three

decades of incredible cognitive science

and uh cognitive neuroscience work in

the field of vision in the world of

vision human vision that has really

established a couple of really critical

northstar problems uh just understanding

human visual processing and human

intelligence and one of them is the

recognition of understanding of natural

objects and natural things because a lot

of the psychology and cognitive science

work is pointing to us that is an

innately optimized whatever that word is

uh functionality and ability of human

intelligence. It's more um robust,

faster and uh more nuanced than we had

thought. We even find neuro cororalates

brain areas devoted to faces or places

or body parts, you know. So, so these

two things led to my PhD study of using

machine learning methods to work on real

world object uh recognition.

But it became very painful very quickly

that we are coming banging against one

of the most um continue to be the most

important um challenge in AI and machine

learning is the lack of generalizable

generalizability.

>> You know you can de design a beautiful

model all you want if you're overfitting

that model.

>> Yeah. I remember when it used to be

possible to publish a computer vision

paper, you know, showing your works on

one image.

>> Exactly.

>> Yeah. It's just it's just the

overfitting the models are not very

expressive and uh and uh we lack the

data and we also as a field was betting

on uh making the variables very express

uh rich by like handgineer features.

remember you know every variable

carrying a ton of semantic meaning but

with hand engineer features. So

>> and then towards the end of my PhD my u

advisor pro and I start to look at each

other and say well boy we need more

data. We if we believe in this northstar

problem of object recognition and we

look back at the tools we have you know

mathematically speaking we're

overfitting every model we're

encountering we need to take a fresh

look at this so one thing led to another

he he and I decided we'll just do a at

that point we think it was a largecale

data project called Caltech 101

>> and

>> I remember the data set I wrote papers

using you know your Caltech 101 you and

your uh early graduate student.

>> It helped benefit a lot of researchers

even that Caltech 101 data.

>> Yeah, that was me and my mom labeling

images.

Oh, and a couple of undergrads. But that

was it was the early days of internet.

So suddenly the availability of data was

a new thing. You suddenly I remember

Petro still have this super expensive

digital camera. Um, I think it was Canon

or something like $6,000 walking around

Caltech taking pictures. But we are the

the the the internet generation. I go to

Google image search, I start to see

these thousands and tens of thousands of

images and I tell pro, let's just

download. Of course, it's not that easy

to download. So, one thing led to

another. We built this Caltech uh 101

data set of 101 object categories and

about I would say 30 to 50 30,000 uh

pictures. I think it's really

interesting that um you know even though

everyone's heard of imageet today um

even you kind of took a couple

iterations where you did Caltech 101 and

that was a success lots of people used

it but it's the even the early learnings

from building Caltech 101 that gave you

the basis to build what turned out to be

even an even bigger success

>> right except that at by the time we

start I became an assistant professor um

we started to look at the problem. I

realized it's way bigger than we think.

Just mathematically speaking, Caltech

101 was not sufficient to to power the

the algorithms. We decided to imitate to

do image that and that that was the time

people start to think we're we're too

we're doing too much, right? It's uh

it's just too crazy. the idea of

downloading the entire internet of

images and mapping out all the English

nouns uh was a little bit uh I I start

to get a lot of push back. I remember at

one of the CVPR conference when I

presented the early idea of image that a

couple of researchers publicly

questioned and say said if you cannot

recognize one category of object let's

say the chair you're sitting in how do

you imagine or what's the use of a data

set of 22,000 classes and 15 million

images.

>> Yeah. Yeah. But but in the end you know

that giant data set unlocked a lot of

value for you know countless number of

researchers around the world. So that so

that works

>> well I I think it it was the combination

of betting on the right northstar

problem

>> and the data that drives it.

>> I see.

>> So it was a fun process.

>> Yeah. And and you know to me one of the

um when I think about that story it it

seems like one of those examples where

you know sometimes people feel like they

should only work on projects that are

the huge thing at the first outset but I

feel like for people working in machine

learning if your first project is a bit

smaller it's totally fine have a good

win use the learnings to build up to

even bigger things and then sometimes

you get a you know imageet size win all

of it.

>> Yeah. Well, but in the meantime, I think

it's also important to be driven by an

audacious goal though. You know, you you

can size your problem or size your

project as local milestones and and and

and so on along this journey. But I also

look at some of our uh current students.

They're so peer pressured by this

current uh climate of publishing

non-stop. It becomes more incremental

papers to just get into a publication

for the sake of it. And uh I I

personally always push my students to

ask the question, what is the northstar

that's driving you?

>> Yeah, that's true. Yeah.

>> And you're right. You know, for myself,

when I do research over the years, I've

always

>> pretty much done what I'm excited about

and where I want to, you know, try to

push the field forward. Doesn't mean

don't listen to people. I had to listen

to people, let them shape your opinion,

but in the end, I think the best

researchers, um, let the world shape

their opinion, but in the end, drive

things forward using their own opinion.

>> Totally agree. Yeah, it's your own inner

fire right?

>> Yeah, I think so. Yeah. So as your

research program developed, you've wound

up taking your let's say foundations in

computer vision and neuroscience and

applying it to all sorts of topics

including you know very visibly

healthcare look at neuroscience

applications. Um would love to hear a

bit more about that.

>> Yeah happy to. I I think u the evolution

of my research in computer vision also

kind of follows the evolution of visual

intelligence in animals and there are

two topics that truly excites me. One is

what is a truly impactful application

area that would help human lives and

that's my healthc care work. The other

one is what is vision at the end of the

day about and that brings me to the um

the the trying to close the loop between

perception and robotic learning. So on

the healthc care side um you know one

thing Andrew there was a number that

shocked me about 10 years ago when I met

my long-term collaborator Dr. Arie

Milstein at Stanford Medical School and

that number is about a quarter of a

million Americans die of medical errors

every year.

I had never imagined a number being that

high due to medical errors. There are

many many reasons but we can rest assure

most of the reasons are not intentional.

These are errors of unintended you know

mistakes and and so on. For example,

>> that's a mindboggling number, right? I

think it is.

>> It's made about 40,000 deaths a year

from uh automotive accidents, which is

completely tragic. And this is even

vastly greater.

>> I was going to say that. I'm glad you

brought it up. Just one example, one

number within that mind-boggling number

is the number of hospitalacquired

infection resulted fatality is more than

95,000.

That is more than that's 2.5 times than

the death of uh car accidents. And uh

and in this particular case hospital

acquired infection is a result of many

things but by in large um uh lack of

good hand hygiene practice. I see.

>> So if you look at WH uh there has been a

lot of protocols about clinicians hand

hygiene practice. But in real health

care delivery, um the the when things

get busy and when the process is tedious

and when there's a lack of feedback

system, you still make a lot of

mistakes. Another uh medical um tragic

medical fact is that more than 70

billion dollars every year are spent in

um in fall resulted injuries and

fatalities. And most of this happened to

elderlys at home, but also in the

hospital rooms. And these are huge

issues. And when Arie and I got together

back in 2012, it was the height of

self-driving car um

let's say not hype, but what's the word,

the right word, excitement in Silicon

Valley. And then we look at the

technology of smart sensing cameras,

lightars um radars whatever smart

sensors, machine learning algorithm and

holistic, uh, understanding of a complex

environment with high stakes for human

lives.

I was looking at all that for

self-driving car and realized in healthc

care delivery, we have the same

situation. Much of the process, the

human behavior process of health care is

in the dark. And if we could have smart

sensors, be it in patient rooms or or

senior homes to help our clinicians and

patients to stay safer, that would be

amazing. So Arie and I embarked on this

what we call ambient intelligence

research agenda. Um but one thing I

learned which probably will lead to our

uh other topics um is as soon as you're

applying AI to real human conditions

there's a lot of human issues in

addition to machine learning issues for

example privacy

>> and I remember reading some of your

papers on with with Arie and found it

really interesting how you could build

and deploy systems that were you know

relatively privacy preserving.

>> Yeah. Well thank you. Well, the first uh

iteration of that technology is we use

cameras that do not capture RGB

information. And you've used a lot of

that in self-driving car, the depth

cameras for example. And there you you

you preserve a lot of uh privacy uh

privacy information just by not seeing

the faces and the identity of the

people. But what's really interesting

over the past decade is the changes of

technology is actually giving us a

bigger tool set for privacy com uh

preserved uh computing in this

condition. For example, um um ondevice

inference. You know, as the chips

getting more and more powerful, if you

don't have to transmit any data through

the network and to the central server,

you help people better. Federated

learning, we know it's still early

stage, but that's another um potential

tool for privacy uh preserve computing

and then differential privacy and uh uh

also encryption technologies. So, we're

starting to see that human demand, you

know, privacy and other issues is

driving actually a new wave of machine

learning technology in in uh ambient

intelligence in healthcare. Yeah, that's

really Yeah, I've been encouraged to see

the, you know, real practical

applications of uh differential privacy

that are actually real.

>> Federated learning as you said, probably

the PR is a little bit ahead of the

reality, but I think we'll get there.

But but it's interesting how consumers

in the last several years have uh

fortunately gotten, you know, much more

knowledgeable about privacy and are

increasingly

>> so important. I think the public is also

making us to be better scientists.

>> See? Yeah. Yeah. And and I think and I

think ultimately you know people

understanding AI holds everyone

including us but holds everyone

accountable

>> totally

>> for really doing the right thing.

>> Yeah.

>> Yeah. And you know, and on that note,

one of the really interesting pieces of

work you've been doing has been um

leading several efforts to help educate

legislators or help governments,

especially US government, um work toward

better laws and better regulation,

especially as it relates to AI. Um that

sounds like very important and and I

suspect some days of the week I would

guess somewhat frustrating work, but we

would love to hear more about that.

Yeah. So I think first of all I have to

credit many many people. So about uh um

four years ago and I was actually

finishing my sbatical from Google time.

I was very privileged to work with so

many businesses you know enterprise

developers just just a large um number

and variety of vertical industries and

realizing AI's human impact. And that

was when many um faculty leaders at

Stanford and also just uh our president

provost former president and former

provos all get together and realize

there is a role historical role that

Stanford needs to uh play in the

advances of AI. We were part of the part

part of the birthplace of AI. you know a

lot of work uh our uh previous um

generation have done and a lot of work

you've done and and some of our work

I've done led to AI today's AI what is

our historical opportunity and

responsibility with that we believe that

the next generation of AI education and

research and policy needs to be human-

centered and having established the

human- center AI institute what we call

hai. One of the work that really took me

outside of my comfort zone or any

expertise is really a uh deeper uh

engagement with policy thinkers and

makers because you know we're here in

Silicon Valley and there is a culture in

Silicon Valley is we just keep making

things and the law will catch up by

itself but AI is impacting human lives

and some sometimes negatively so rapidly

that it is not good for any of us if we

the experts are not at the table with

the policy thinkers and makers to really

try to make this technology better for

the people. I mean we're talking about

fairness, we're talking about privacy.

uh we also are talking about the brain

drain of AI to industry and the the um

uh concentration of data and compute in

a small number of uh of technology

companies. All this are really part of

the changes of our time. Some are really

exciting changes, some have profound

impact that we don't we cannot

necessarily uh predict yet. So one of

the policy work that Stanford HAI has

very proudly engaged in is we were the

one of the leading universities that

lobbied a bill called the national AI

research cloud task force bill. It

changed the name from research cloud to

research resource. So now the bill's

acronym is NAR national AI research

resource. And this bill is calling for a

task force uh to put together a roadmap

for America's public sector especially

higher education and research um uh

sector to um increase their access to

resource for AI compute and AI data and

it really is aimed to rejuvenate

America's uh uh ecosystem in in AI

innovation and research and I'm on the

12 person task force uh for under Biden

administration for this bill. And we

hope that's a piece of policy that is

not a regulatory policy. It's more an

incentive policy to build and rejuvenate

ecosystems.

>> I'm glad that you're doing this to help

shape US policy and this type of making

sure enough resources are allocated to,

you know, ensure healthy development of

AI. It feels like this is something that

um every country needs at this point.

Yeah.

>> So, you know, just from the things that

you are doing by yourself, not to speak

of the things that the global AI

community is doing, there's just so much

going on in AI right now. So many

opportunities, so much excitement.

>> Um,

>> I found that for someone getting started

in machine learning for the first time,

sometimes there's so much going on, it

can almost feel a little bit

overwhelming.

>> Totally. What advice do you have for

someone getting started in machine

learning?

>> Great question, Andrew. I'm sure you

have great advice. You're one of the

well-known advocate for AI machine

learning education. So, um I do get this

question a lot as well. And one thing

you're totally right is AI really today

feels different from our time. During

our

>> just for the record, now is still our

time.

Yeah, that's true. When we were starting

in AI, I love that. Exactly. We're still

part of this. Um, when we get started,

the entrance to AI and machine learning

was relatively narrow. You almost have

to start from computer science angle,

right? you know, as a physics major, I

still had to wedge myself into the

computer science track or electrical

engineering track to to get to AI. But

today, I actually think that there is

many aspect of uh AI that that creates

entry points for people from all walks

of life. Uh on the technical side, I

think uh it's obvious that there's just

a incredible plethora of resources out

there on the internet from Corsera to

you know YouTube to Tik Tok to GitHub to

the to there's just so much that

students worldwide can u learn about AI

and machine learning compared to the

time we began learning machine learning

and also any campuses we're not talking

about just college campuses, we're

talking about high school campuses or

even sometimes uh earlier. Uh we're

starting to see more um available um

classes and resources. So I think there

is I do encourage those of the young

people with a technical interest and um

resource and opportunity to embrace

these uh resources because it's it's a

lot of fun. But having said that, for

those of you who are not coming from a

technical angle, who still are

passionate about AI, whether it's the

downstream application or the the

creativity it engenders or the policy

and social angle or important social

problems whether it's e digital

economics or the governance governance

or you know history, ethics, uh

political sciences there. I do invite

you to to join us because there is a lot

of work to be done. There is a lot of uh

a lot of unknown questions. For example,

my colleague at hi are questioning are

are trying to find answers on how do you

define our economy in the digital age?

What does it mean when robots software

are participating in the workflow more

and more? How do you measure our

economy? That that's not a AI coding

question. That is a AI impact question.

>> We're looking at the incredible advances

of generative AI and there will be more.

Um what does that mean for creative

creativity and to the creators from

music to art to uh writing? Um I think

you know there is a a lot of concerns

and I think it's rightfully so. But in

the meantime, um, it takes people

together to figure this out and and also

to use this new tool.

>> So I think in short, I just think it's a

it's a very exciting time and anybody

with any walks of life, as long as

you're passionate about this, there's a

role to play.

>> Yeah, I think that's really exciting.

When we talk about economics, think

about, you know, my conversations with

um, Professor Eric Brenoffson, right?

studying

>> impact of AI on the economy. But

>> from from what you're saying and and I

agree, it seems like no matter what your

current interests are,

>> AI is such a general purpose technology

that the combination of your current

interest and AI is often promising.

Yeah.

>> And I find that even for learners that

may not yet have a specific interest,

you know, if you find your way into AI,

start learning things, often the

interest will evolve and then you can

start to craft your own path.

>> Y

>> and given where AI is today, there's

still so much room and so much need for

a lot more people to craft their own

paths to do this exciting work that I

think the world still needs a lot more

of.

>> Totally agree. Yeah. So one piece of

work that you did I thought was very

cool was starting a program initially

called sailors and then later AI for all

which was really reaching out to you

know high school and even younger

students to try to give them more

opportunities in AI including people of

all walks of life. Love to hear more

about that. Yeah. Well, this is in the

spirit of this conversation is uh um

that was back in 2015. Um that was uh

there was starting to be a lot of

excitement of AI but there was also

starting to be this talk about

killer robot coming next door

terminators coming and I I was at that

time Andrew um I was the director of

Stephi lab and I was thinking

you know we know how far we are from

terminators coming and that seemed to be

a really a little bit of far-fetched

concern term but I was living my work

life with a real concern I felt no one

was talking about which was the lack of

representation in AI at that time I

guess after Daphne has left I was the

only woman faculty at Stanford AI lab I

see

>> and uh and we're having very small

around 15% of uh women graduate students

and we really don't see anybody from uh

the underrepresented minority groups in

Stanford AI um program and this is a

national or even worldwide issue. So it

wasn't just Stanford.

>> Frankly, it still needs a lot of work

today.

>> Exactly. So how do we do this? Well, I

got together with my former student Olga

Rousovski and also a uh um a long-term

educator of STEM uh topics uh Dr. Rick

Summer from a Stanford pre-colgiate

study program and thought about inviting

high schoolers at that time, women, high

school young women to participate in a

summer program to inspire them to to

learn AI and and that was how it started

in 2015 and 2017. We got um a lot of

encouragement and support from people

like Jensen and Lorie Hang and Melinda

Gates and we uh formed the national

nonprofit called AI for all which is

really committed to training or or or or

uh helping tomorrow's leaders uh shaping

tomorrow's leaders for AI uh from

students of all walks of life especially

the traditionally underserved and

underrepresented communities.

And uh uh we you know till today we've

had many many summer camps and summer

programs across the country. More than

uh um um more than 15 universities are

involved and uh uh we have online

curriculum to encourage students as well

as college pathway programs to to

continue support these students uh uh

career by matching them with internships

and mentors. So it's a it's a continued

effort of encouraging students of all

walks of life.

>> Yeah. And I remember back then, I think

your group was printing these really

cool t-shirts that asked the question,

um, AI will change the world. Who will

change AI? And I thought the answer of

making sure everyone can come in and

participate, that was a great answer.

>> Yeah. Still an important question today.

>> Yeah. So, that's a great thought and I

think that takes us toward the end of

the interview. Um, any final thoughts

for the people watching this? still that

um this is a very nent field as you said

Andrew we are still in the middle of

this I I still feel there's just so many

questions that uh you know I wake up

excited to work on with my students in

the lab and I think uh there's a lot

more opportunities for for the young

people out there who want to learn and

contribute and shape tomorrow's AI

>> yeah well said that's that's very

inspiring really great to chat to you

and thank you for

>> Thank you. It's fun to have these

conversations.

[Music]