[MUSIC] Okay, given your great historical
perspective of machine learning- >> Are you calling me old? >> I am calling you old. >> It's always like this. >> [LAUGH] I was wondering if you
could give me a little perspective on the future. >> Well, let's talk about the present. >> Okay. >> [LAUGH]
>> I'll accept that, we can talk about the present. He always makes things really long and
really detailed. So just talk about the present. We'll get to the future eventually,
don't worry. >> [LAUGH]
>> Here we go, the present. What is the state of machine learning,
Carlos? >> [CROSSTALK] I'm really excited about
machine learning today, honestly. Being old, that you work in
machine learning for a long time. But today, every time I open a web app,
it uses machine learning inside it. And that's pretty cool. And I pull out my phone, where's my phone? You always have your phone with you. >> I actually don't have
my phone right now. >> She's always on her phone. Anyway, when you pull out your phone,
every time you use it, there's a system that uses machine learning
right inside the cloud that supports it. Which is really pretty cool. And now with devices, check this out. I got this watch. >> Pretty spiffy.
>> It's a wonderful watch that tracks my activity,
knows when I'm running, when I'm biking- >> Knows when he's sleeping or not sleeping. Always this debate,
who got less sleep with the baby? [LAUGH]
>> [LAUGH] So, yeah, that's a wonderful watch. And underneath these things,
it's all using machine learning. So we see the power of machine
learning just kind of, again and again, being used today. But I think we've only scratched the
surface of what machine learning can do. Where do you see machine
learning going in the future? You tell me. >> Where do I see it? Well, I know probably a lot of ways in
which it would be useful to Carlos. One being the fact that he hates driving. So of course, there is this smart car idea
that this self-driving, not smart car. >> They're also smart. >> Are they smart? They're pretty smart, I guess. It's a self-driving car. A smart car is a brand. That's not what I mean,
I don't mean those little cars. I mean normal cars [LAUGH]
that drive themselves. >> That would be really useful for
me, because I don't like driving. I only like biking. And so I need some-
>> Self-driving bikes. >> No, no, I like driving my bike. >> [CROSSTALK] That's for
when you get really, really old. [LAUGH]
>> [CROSSTALK] So I'm gonna talk about that. But when I think about me getting old,
I think about how machine learning can really have impact with
your personalized medicine. And the question here is,
why is it when I go to a doctor for certain disease or certain condition, the treatment that Emily gets is exactly
the same as the treatment that I get. I mean, we're different people
with different lifestyles, with different body types, as you can see. And still, with medicine,
we're getting exactly the same treatment. That makes no sense. The treatment needs to be
personalized to who I am, what I do, how much of a couch potato I am,
what my DNA looks like. And with machine learning, we're going
to be able to do that all the time. It's really super cool,
I'm pretty excited. >> That was just so
optimistic, which is good. It's really nice, but it's hard! It's really, really hard to do this. The data is really noisy, cuz there's some
doctor who's there taking these records, making notes about Carlos' health
conditions compared to mine. And then there's lots of missing data,
because maybe Carlos skipped his last checkup or
the doctor forgot to enter something. Or maybe the doctor put something under
a different entry than they did for me. And so somehow our machine learning
algorithms are gonna have to be able to parse this really complex and noisy data, in order to do this
type of personalized medicine. Which I think is a possibility. I think it's really cool, but-
>> It is hard. All these other things we're
talking about, from self-driving cars which we're seeing as a reality
today, to personalized medicine. It's all hard stuff. But in the future, and
this is the optimistic in me, the world is going to
be personalized to me. >> [LAUGH]
>> You know what I mean? >> The grand goal of everything for
Carlos. [LAUGH]
>> [LAUGH] For me and for Emily, too. She can have her own world. So the idea that today
when we go to the Web and we buy things, we have to go through
very similar experiences as others. But as the world becomes smarter, as more
data is collected, as we come up with this algorithms which are robust to all
this kind of noise in the data, we're gonna have experiences that
are really unique to us, idiosyncratic. And we're not gonna suffer
as much from technology. Technology's gonna really adapt to us,
which is really super cool. >> I agree, I think that is a really,
really exciting goal. And I think that is one of the directions
that machine learning is going in. But I think something else that's really
cool is even doing the things we can do right now. But doing them as the data's coming in,
in real time, with really complicated different types of data sets, and
putting all this information together. And maybe it's just even
still recommending products, something we've done before in the past,
but with different types of data. And again, in this really fast,
rapid, online type of setting, and getting feedback and incorporating this
to make these types of prediction, yeah. >> Yeah, dealing with temporal data,
which is the topic here, as well as the spatial
data like location data. All this different kind of data sets,
images, video, audio, sensing data, combining all these different data sources
in real time is one of the big challenges. And what I'm excited to see is that
there's a lot more machine learning work and a lot more algorithms being developed
that try to address this confusion of different data types. And that's a really cool
topic to work on right now. >> For sure, so. >> Of course. >> We hope that's really exciting to you,
in addition to us. >> Yeah, it is exciting. And of course, the methods that
we develop have to scale to these massive data sets that we're talking
about, different kinds of data. Be able to be really useful and
robust as you are analyzing the data. And honestly, it shouldn't require a PhD in machine learning to be able to
access the systems and be able to use it. So part of the challenge that
we have to do is be able to make these algorithms robust and
these methods robust. So that a wide range of people,
even those that don't have the most theoretical background
are able to build the smartest, coolest applications
using machine learning. That's exciting to me as well. >> Absolutely. [MUSIC]