WEBVTT
Kind: captions
Language: en

00:00:00.000 --> 00:00:16.400 align:start position:0%
 
[Music]

00:00:16.400 --> 00:00:16.410 align:start position:0%
 
 

00:00:16.410 --> 00:00:20.400 align:start position:0%
 
[Music]

00:00:20.400 --> 00:00:20.410 align:start position:0%
 
 

00:00:20.410 --> 00:00:24.100 align:start position:0%
 
so<00:00:21.550><c> AI</c><00:00:22.550><c> has</c><00:00:22.970><c> gone</c><00:00:23.240><c> through</c><00:00:23.420><c> a</c><00:00:23.480><c> little</c><00:00:23.840><c> bit</c><00:00:23.960><c> of</c>

00:00:24.100 --> 00:00:24.110 align:start position:0%
so AI has gone through a little bit of
 

00:00:24.110 --> 00:00:26.970 align:start position:0%
so AI has gone through a little bit of
it<00:00:24.259><c> you</c><00:00:24.500><c> should</c><00:00:24.560><c> usage</c><00:00:25.689><c> transformation</c><00:00:26.689><c> last</c>

00:00:26.970 --> 00:00:26.980 align:start position:0%
it you should usage transformation last
 

00:00:26.980 --> 00:00:30.790 align:start position:0%
it you should usage transformation last
12<00:00:27.980><c> months</c><00:00:28.750><c> historically</c><00:00:29.750><c> I</c><00:00:30.080><c> think</c><00:00:30.529><c> it's</c><00:00:30.619><c> been</c>

00:00:30.790 --> 00:00:30.800 align:start position:0%
12 months historically I think it's been
 

00:00:30.800 --> 00:00:35.530 align:start position:0%
12 months historically I think it's been
integrated<00:00:31.700><c> AI</c><00:00:31.820><c> that</c><00:00:32.180><c> means</c><00:00:32.419><c> a</c><00:00:34.360><c> machine</c><00:00:35.360><c> that</c>

00:00:35.530 --> 00:00:35.540 align:start position:0%
integrated AI that means a machine that
 

00:00:35.540 --> 00:00:39.180 align:start position:0%
integrated AI that means a machine that
can<00:00:35.660><c> do</c><00:00:35.950><c> several</c><00:00:36.950><c> different</c><00:00:37.690><c> human-like</c>

00:00:39.180 --> 00:00:39.190 align:start position:0%
can do several different human-like
 

00:00:39.190 --> 00:00:42.069 align:start position:0%
can do several different human-like
things<00:00:40.190><c> tasks</c><00:00:40.790><c> and</c><00:00:40.940><c> it</c><00:00:41.120><c> has</c><00:00:41.690><c> those</c><00:00:41.840><c> abilities</c>

00:00:42.069 --> 00:00:42.079 align:start position:0%
things tasks and it has those abilities
 

00:00:42.079 --> 00:00:45.930 align:start position:0%
things tasks and it has those abilities
so<00:00:43.510><c> being</c><00:00:44.510><c> able</c><00:00:44.600><c> to</c><00:00:44.719><c> chat</c><00:00:44.930><c> being</c><00:00:45.200><c> one</c><00:00:45.350><c> of</c><00:00:45.380><c> them</c>

00:00:45.930 --> 00:00:45.940 align:start position:0%
so being able to chat being one of them
 

00:00:45.940 --> 00:00:48.490 align:start position:0%
so being able to chat being one of them
until<00:00:46.940><c> recently</c><00:00:47.360><c> I</c><00:00:47.450><c> think</c><00:00:47.600><c> I</c><00:00:47.989><c> have</c><00:00:48.140><c> thought</c><00:00:48.350><c> it</c>

00:00:48.490 --> 00:00:48.500 align:start position:0%
until recently I think I have thought it
 

00:00:48.500 --> 00:00:51.189 align:start position:0%
until recently I think I have thought it
as<00:00:48.980><c> I've</c><00:00:49.910><c> thought</c><00:00:49.969><c> of</c><00:00:50.329><c> it</c><00:00:50.449><c> in</c><00:00:50.539><c> terms</c><00:00:50.719><c> of</c><00:00:50.929><c> deep</c>

00:00:51.189 --> 00:00:51.199 align:start position:0%
as I've thought of it in terms of deep
 

00:00:51.199 --> 00:00:53.079 align:start position:0%
as I've thought of it in terms of deep
learning<00:00:51.350><c> and</c><00:00:51.679><c> the</c><00:00:52.489><c> reason</c><00:00:52.699><c> for</c><00:00:52.789><c> that</c><00:00:52.910><c> is</c><00:00:53.059><c> that</c>

00:00:53.079 --> 00:00:53.089 align:start position:0%
learning and the reason for that is that
 

00:00:53.089 --> 00:00:55.000 align:start position:0%
learning and the reason for that is that
deep<00:00:53.420><c> learning</c><00:00:53.570><c> has</c><00:00:53.960><c> really</c><00:00:54.350><c> helped</c><00:00:54.739><c> crack</c>

00:00:55.000 --> 00:00:55.010 align:start position:0%
deep learning has really helped crack
 

00:00:55.010 --> 00:00:57.000 align:start position:0%
deep learning has really helped crack
open<00:00:55.489><c> a</c><00:00:55.579><c> lot</c><00:00:55.730><c> of</c><00:00:55.820><c> those</c><00:00:55.910><c> capabilities</c>

00:00:57.000 --> 00:00:57.010 align:start position:0%
open a lot of those capabilities
 

00:00:57.010 --> 00:00:59.590 align:start position:0%
open a lot of those capabilities
nowadays<00:00:58.010><c> it's</c><00:00:58.579><c> becoming</c><00:00:58.999><c> quickly</c>

00:00:59.590 --> 00:00:59.600 align:start position:0%
nowadays it's becoming quickly
 

00:00:59.600 --> 00:01:08.020 align:start position:0%
nowadays it's becoming quickly
synonymous<00:01:00.140><c> with</c><00:01:00.289><c> just</c><00:01:00.499><c> machinery</c><00:01:04.870><c> so</c><00:01:07.030><c> in</c>

00:01:08.020 --> 00:01:08.030 align:start position:0%
synonymous with just machinery so in
 

00:01:08.030 --> 00:01:11.170 align:start position:0%
synonymous with just machinery so in
order<00:01:08.210><c> to</c><00:01:08.330><c> an</c><00:01:08.540><c> integrated</c><00:01:09.020><c> AI</c><00:01:09.110><c> you</c><00:01:10.070><c> to</c><00:01:10.970><c> create</c>

00:01:11.170 --> 00:01:11.180 align:start position:0%
order to an integrated AI you to create
 

00:01:11.180 --> 00:01:13.570 align:start position:0%
order to an integrated AI you to create
a<00:01:11.900><c> machine</c><00:01:12.530><c> that</c><00:01:12.830><c> can</c><00:01:12.920><c> do</c><00:01:13.070><c> a</c><00:01:13.100><c> lot</c><00:01:13.340><c> of</c><00:01:13.370><c> the</c>

00:01:13.570 --> 00:01:13.580 align:start position:0%
a machine that can do a lot of the
 

00:01:13.580 --> 00:01:15.000 align:start position:0%
a machine that can do a lot of the
things<00:01:13.790><c> core</c><00:01:14.150><c> things</c><00:01:14.360><c> that</c><00:01:14.420><c> human</c><00:01:14.750><c> could</c><00:01:14.840><c> do</c>

00:01:15.000 --> 00:01:15.010 align:start position:0%
things core things that human could do
 

00:01:15.010 --> 00:01:18.190 align:start position:0%
things core things that human could do
you<00:01:16.010><c> would</c><00:01:16.130><c> need</c><00:01:16.340><c> to</c><00:01:17.090><c> be</c><00:01:17.210><c> able</c><00:01:17.300><c> to</c><00:01:17.479><c> sense</c>

00:01:18.190 --> 00:01:18.200 align:start position:0%
you would need to be able to sense
 

00:01:18.200 --> 00:01:19.930 align:start position:0%
you would need to be able to sense
things<00:01:18.380><c> so</c><00:01:18.650><c> maybe</c><00:01:18.770><c> look</c><00:01:19.490><c> at</c><00:01:19.580><c> images</c><00:01:19.880><c> and</c>

00:01:19.930 --> 00:01:19.940 align:start position:0%
things so maybe look at images and
 

00:01:19.940 --> 00:01:22.420 align:start position:0%
things so maybe look at images and
recognize<00:01:20.390><c> them</c><00:01:20.600><c> you'd</c><00:01:21.200><c> be</c><00:01:21.350><c> able</c><00:01:21.440><c> to</c><00:01:21.560><c> to</c><00:01:22.220><c> hear</c>

00:01:22.420 --> 00:01:22.430 align:start position:0%
recognize them you'd be able to to hear
 

00:01:22.430 --> 00:01:27.070 align:start position:0%
recognize them you'd be able to to hear
things<00:01:22.720><c> so</c><00:01:23.720><c> so</c><00:01:24.730><c> voice</c><00:01:25.730><c> voice</c><00:01:26.330><c> to</c><00:01:26.570><c> speech</c><00:01:26.810><c> or</c>

00:01:27.070 --> 00:01:27.080 align:start position:0%
things so so voice voice to speech or
 

00:01:27.080 --> 00:01:31.750 align:start position:0%
things so so voice voice to speech or
speech<00:01:27.800><c> to</c><00:01:28.100><c> text</c><00:01:28.960><c> and</c><00:01:29.960><c> maybe</c><00:01:30.710><c> you</c><00:01:30.920><c> need</c><00:01:31.490><c> to</c><00:01:31.640><c> be</c>

00:01:31.750 --> 00:01:31.760 align:start position:0%
speech to text and maybe you need to be
 

00:01:31.760 --> 00:01:34.390 align:start position:0%
speech to text and maybe you need to be
able<00:01:31.850><c> to</c><00:01:32.380><c> comprehend</c><00:01:33.380><c> so</c><00:01:33.680><c> natural</c><00:01:34.190><c> language</c>

00:01:34.390 --> 00:01:34.400 align:start position:0%
able to comprehend so natural language
 

00:01:34.400 --> 00:01:38.140 align:start position:0%
able to comprehend so natural language
understanding<00:01:35.380><c> and</c><00:01:36.380><c> you</c><00:01:36.680><c> may</c><00:01:37.190><c> even</c><00:01:37.400><c> need</c><00:01:38.000><c> to</c>

00:01:38.140 --> 00:01:38.150 align:start position:0%
understanding and you may even need to
 

00:01:38.150 --> 00:01:41.050 align:start position:0%
understanding and you may even need to
be<00:01:38.270><c> able</c><00:01:38.360><c> to</c><00:01:38.590><c> physically</c><00:01:39.590><c> interact</c><00:01:40.060><c> there</c>

00:01:41.050 --> 00:01:41.060 align:start position:0%
be able to physically interact there
 

00:01:41.060 --> 00:01:43.030 align:start position:0%
be able to physically interact there
have<00:01:41.180><c> been</c><00:01:41.300><c> huge</c><00:01:41.870><c> advances</c><00:01:42.380><c> in</c><00:01:42.500><c> each</c><00:01:42.680><c> of</c><00:01:42.860><c> those</c>

00:01:43.030 --> 00:01:43.040 align:start position:0%
have been huge advances in each of those
 

00:01:43.040 --> 00:01:47.140 align:start position:0%
have been huge advances in each of those
directions<00:01:43.510><c> recently</c><00:01:44.800><c> so</c><00:01:45.800><c> Q</c><00:01:46.400><c> learning</c><00:01:46.520><c> for</c>

00:01:47.140 --> 00:01:47.150 align:start position:0%
directions recently so Q learning for
 

00:01:47.150 --> 00:01:49.630 align:start position:0%
directions recently so Q learning for
human<00:01:47.540><c> interaction</c><00:01:48.430><c> combinations</c><00:01:49.430><c> of</c>

00:01:49.630 --> 00:01:49.640 align:start position:0%
human interaction combinations of
 

00:01:49.640 --> 00:01:52.180 align:start position:0%
human interaction combinations of
advanced<00:01:50.270><c> neural</c><00:01:50.450><c> embedding</c><00:01:50.930><c> for</c><00:01:51.350><c> natural</c>

00:01:52.180 --> 00:01:52.190 align:start position:0%
advanced neural embedding for natural
 

00:01:52.190 --> 00:01:54.190 align:start position:0%
advanced neural embedding for natural
language<00:01:52.280><c> understanding</c><00:01:53.200><c> convolutional</c>

00:01:54.190 --> 00:01:54.200 align:start position:0%
language understanding convolutional
 

00:01:54.200 --> 00:01:58.690 align:start position:0%
language understanding convolutional
neural<00:01:54.590><c> nets</c><00:01:54.860><c> have</c><00:01:55.120><c> really</c><00:01:56.120><c> driven</c><00:01:57.110><c> the</c><00:01:57.700><c> image</c>

00:01:58.690 --> 00:01:58.700 align:start position:0%
neural nets have really driven the image
 

00:01:58.700 --> 00:02:01.720 align:start position:0%
neural nets have really driven the image
recognition<00:01:59.800><c> pretty</c><00:02:00.800><c> significantly</c><00:02:01.370><c> for</c><00:02:01.580><c> the</c>

00:02:01.720 --> 00:02:01.730 align:start position:0%
recognition pretty significantly for the
 

00:02:01.730 --> 00:02:04.690 align:start position:0%
recognition pretty significantly for the
last<00:02:01.880><c> several</c><00:02:02.060><c> years</c><00:02:02.470><c> and</c><00:02:03.470><c> for</c><00:02:04.340><c> 1d</c>

00:02:04.690 --> 00:02:04.700 align:start position:0%
last several years and for 1d
 

00:02:04.700 --> 00:02:07.360 align:start position:0%
last several years and for 1d
convolutions<00:02:05.480><c> also</c><00:02:06.020><c> especially</c><00:02:06.590><c> some</c><00:02:07.160><c> work</c>

00:02:07.360 --> 00:02:07.370 align:start position:0%
convolutions also especially some work
 

00:02:07.370 --> 00:02:10.809 align:start position:0%
convolutions also especially some work
that<00:02:07.520><c> Baidu's</c><00:02:07.880><c> been</c><00:02:08.030><c> doing</c><00:02:08.179><c> recently</c><00:02:08.530><c> for</c><00:02:09.819><c> for</c>

00:02:10.809 --> 00:02:10.819 align:start position:0%
that Baidu's been doing recently for for
 

00:02:10.819 --> 00:02:14.130 align:start position:0%
that Baidu's been doing recently for for
speech<00:02:11.360><c> to</c><00:02:11.569><c> text</c><00:02:11.750><c> so</c><00:02:12.430><c> we've</c><00:02:13.430><c> made</c><00:02:13.700><c> a</c><00:02:13.760><c> lot</c><00:02:13.879><c> of</c>

00:02:14.130 --> 00:02:14.140 align:start position:0%
speech to text so we've made a lot of
 

00:02:14.140 --> 00:02:16.650 align:start position:0%
speech to text so we've made a lot of
advances<00:02:15.140><c> in</c><00:02:15.170><c> all</c><00:02:15.800><c> of</c><00:02:16.040><c> those</c><00:02:16.129><c> directions</c>

00:02:16.650 --> 00:02:16.660 align:start position:0%
advances in all of those directions
 

00:02:16.660 --> 00:02:19.780 align:start position:0%
advances in all of those directions
nothing's<00:02:17.660><c> perfect</c><00:02:17.840><c> editors</c><00:02:18.430><c> there</c><00:02:19.430><c> I</c><00:02:19.459><c> least</c>

00:02:19.780 --> 00:02:19.790 align:start position:0%
nothing's perfect editors there I least
 

00:02:19.790 --> 00:02:22.780 align:start position:0%
nothing's perfect editors there I least
I<00:02:19.879><c> haven't</c><00:02:20.120><c> seen</c><00:02:20.269><c> any</c><00:02:20.420><c> good</c><00:02:20.629><c> integrated</c>

00:02:22.780 --> 00:02:22.790 align:start position:0%
I haven't seen any good integrated
 

00:02:22.790 --> 00:02:24.910 align:start position:0%
I haven't seen any good integrated
a<00:02:23.329><c> project</c><00:02:24.319><c> that</c><00:02:24.409><c> do</c><00:02:24.530><c> all</c><00:02:24.709><c> of</c><00:02:24.739><c> those</c>

00:02:24.910 --> 00:02:24.920 align:start position:0%
a project that do all of those
 

00:02:24.920 --> 00:02:35.349 align:start position:0%
a project that do all of those
simultaneously<00:02:31.659><c> so</c><00:02:32.659><c> I</c><00:02:33.250><c> I</c><00:02:34.250><c> work</c><00:02:34.939><c> in</c><00:02:35.120><c> several</c>

00:02:35.349 --> 00:02:35.359 align:start position:0%
simultaneously so I I work in several
 

00:02:35.359 --> 00:02:38.709 align:start position:0%
simultaneously so I I work in several
different<00:02:35.599><c> areas</c><00:02:35.859><c> you</c><00:02:36.859><c> know</c><00:02:36.950><c> intact</c><00:02:37.459><c> we</c><00:02:38.120><c> do</c>

00:02:38.709 --> 00:02:38.719 align:start position:0%
different areas you know intact we do
 

00:02:38.719 --> 00:02:42.460 align:start position:0%
different areas you know intact we do
the<00:02:39.439><c> backend</c><00:02:39.890><c> AI</c><00:02:40.459><c> for</c><00:02:41.140><c> for</c><00:02:42.140><c> large</c><00:02:42.290><c> enterprises</c>

00:02:42.460 --> 00:02:42.470 align:start position:0%
the backend AI for for large enterprises
 

00:02:42.470 --> 00:02:45.539 align:start position:0%
the backend AI for for large enterprises
so<00:02:43.750><c> using</c><00:02:44.750><c> neural</c><00:02:45.019><c> networks</c><00:02:45.379><c> for</c>

00:02:45.539 --> 00:02:45.549 align:start position:0%
so using neural networks for
 

00:02:45.549 --> 00:02:47.259 align:start position:0%
so using neural networks for
recommenders<00:02:46.549><c> for</c><00:02:46.909><c> you</c><00:02:47.000><c> as</c><00:02:47.150><c> our</c>

00:02:47.259 --> 00:02:47.269 align:start position:0%
recommenders for you as our
 

00:02:47.269 --> 00:02:49.949 align:start position:0%
recommenders for you as our
understanding<00:02:47.959><c> for</c><00:02:48.379><c> making</c><00:02:49.099><c> sure</c><00:02:49.310><c> that</c><00:02:49.340><c> your</c>

00:02:49.949 --> 00:02:49.959 align:start position:0%
understanding for making sure that your
 

00:02:49.959 --> 00:02:54.069 align:start position:0%
understanding for making sure that your
that<00:02:50.959><c> a</c><00:02:51.200><c> company</c><00:02:51.920><c> can</c><00:02:52.189><c> interact</c><00:02:52.730><c> with</c><00:02:52.790><c> a</c><00:02:53.079><c> with</c>

00:02:54.069 --> 00:02:54.079 align:start position:0%
that a company can interact with a with
 

00:02:54.079 --> 00:02:56.800 align:start position:0%
that a company can interact with a with
a<00:02:54.109><c> a</c><00:02:54.469><c> customer</c><00:02:55.310><c> the</c><00:02:55.400><c> right</c><00:02:55.549><c> way</c><00:02:55.700><c> so</c><00:02:55.970><c> not</c><00:02:56.599><c> just</c>

00:02:56.800 --> 00:02:56.810 align:start position:0%
a a customer the right way so not just
 

00:02:56.810 --> 00:02:58.390 align:start position:0%
a a customer the right way so not just
trying<00:02:56.959><c> to</c><00:02:57.139><c> get</c><00:02:57.379><c> that</c><00:02:57.500><c> customer</c><00:02:58.040><c> to</c><00:02:58.310><c> buy</c>

00:02:58.390 --> 00:02:58.400 align:start position:0%
trying to get that customer to buy
 

00:02:58.400 --> 00:03:00.909 align:start position:0%
trying to get that customer to buy
another<00:02:58.609><c> bag</c><00:02:58.790><c> of</c><00:02:58.909><c> chips</c><00:02:59.060><c> but</c><00:02:59.359><c> creating</c><00:03:00.319><c> a</c><00:03:00.590><c> more</c>

00:03:00.909 --> 00:03:00.919 align:start position:0%
another bag of chips but creating a more
 

00:03:00.919 --> 00:03:03.220 align:start position:0%
another bag of chips but creating a more
long-term<00:03:01.780><c> reinforcement</c><00:03:02.780><c> learning</c><00:03:02.870><c> type</c>

00:03:03.220 --> 00:03:03.230 align:start position:0%
long-term reinforcement learning type
 

00:03:03.230 --> 00:03:05.949 align:start position:0%
long-term reinforcement learning type
relationship<00:03:03.919><c> is</c><00:03:04.849><c> where</c><00:03:05.150><c> a</c><00:03:05.450><c> lot</c><00:03:05.629><c> of</c><00:03:05.659><c> my</c><00:03:05.810><c> focus</c>

00:03:05.949 --> 00:03:05.959 align:start position:0%
relationship is where a lot of my focus
 

00:03:05.959 --> 00:03:14.830 align:start position:0%
relationship is where a lot of my focus
has<00:03:06.230><c> been</c><00:03:06.849><c> recently</c><00:03:12.340><c> so</c><00:03:13.340><c> so</c><00:03:14.030><c> the</c><00:03:14.510><c> by</c><00:03:14.629><c> the</c><00:03:14.689><c> bay</c>

00:03:14.830 --> 00:03:14.840 align:start position:0%
has been recently so so the by the bay
 

00:03:14.840 --> 00:03:16.809 align:start position:0%
has been recently so so the by the bay
conferences<00:03:15.319><c> are</c><00:03:15.409><c> great</c><00:03:15.590><c> I</c><00:03:15.819><c> love</c>

00:03:16.809 --> 00:03:16.819 align:start position:0%
conferences are great I love
 

00:03:16.819 --> 00:03:19.240 align:start position:0%
conferences are great I love
participating<00:03:17.480><c> in</c><00:03:17.569><c> these</c><00:03:17.689><c> and</c><00:03:18.319><c> I</c><00:03:18.709><c> always</c>

00:03:19.240 --> 00:03:19.250 align:start position:0%
participating in these and I always
 

00:03:19.250 --> 00:03:23.319 align:start position:0%
participating in these and I always
learn<00:03:20.060><c> something</c><00:03:20.209><c> new</c><00:03:20.359><c> I</c><00:03:20.689><c> see</c><00:03:21.260><c> I</c><00:03:21.650><c> see</c><00:03:22.459><c> folk</c>

00:03:23.319 --> 00:03:23.329 align:start position:0%
learn something new I see I see folk
 

00:03:23.329 --> 00:03:25.539 align:start position:0%
learn something new I see I see folk
that<00:03:23.569><c> that</c><00:03:24.109><c> that</c><00:03:24.470><c> I</c><00:03:24.500><c> enjoy</c><00:03:24.889><c> meeting</c><00:03:25.459><c> with</c>

00:03:25.539 --> 00:03:25.549 align:start position:0%
that that that I enjoy meeting with
 

00:03:25.549 --> 00:03:29.170 align:start position:0%
that that that I enjoy meeting with
regularly<00:03:25.970><c> and</c><00:03:26.590><c> the</c><00:03:27.590><c> Lexx</c><00:03:27.769><c> is</c><00:03:27.889><c> great</c><00:03:28.099><c> also</c>

00:03:29.170 --> 00:03:29.180 align:start position:0%
regularly and the Lexx is great also
 

00:03:29.180 --> 00:03:34.189 align:start position:0%
regularly and the Lexx is great also
[Music]