WEBVTT
Kind: captions
Language: en

00:00:01.399 --> 00:00:04.340 align:start position:0%
 
our<00:00:02.399><c> ability</c><00:00:02.700><c> to</c><00:00:02.970><c> learn</c><00:00:03.270><c> and</c><00:00:03.629><c> get</c><00:00:03.899><c> better</c><00:00:03.959><c> at</c>

00:00:04.340 --> 00:00:04.350 align:start position:0%
our ability to learn and get better at
 

00:00:04.350 --> 00:00:06.320 align:start position:0%
our ability to learn and get better at
tasks<00:00:04.589><c> through</c><00:00:05.009><c> experience</c><00:00:05.670><c> is</c><00:00:05.879><c> part</c><00:00:06.240><c> of</c>

00:00:06.320 --> 00:00:06.330 align:start position:0%
tasks through experience is part of
 

00:00:06.330 --> 00:00:08.660 align:start position:0%
tasks through experience is part of
being<00:00:06.509><c> human</c><00:00:06.750><c> when</c><00:00:07.620><c> we're</c><00:00:07.830><c> bored</c><00:00:08.069><c> we</c><00:00:08.250><c> know</c>

00:00:08.660 --> 00:00:08.670 align:start position:0%
being human when we're bored we know
 

00:00:08.670 --> 00:00:11.000 align:start position:0%
being human when we're bored we know
almost<00:00:09.030><c> nothing</c><00:00:09.330><c> and</c><00:00:09.809><c> can</c><00:00:09.990><c> do</c><00:00:10.200><c> almost</c><00:00:10.500><c> nothing</c>

00:00:11.000 --> 00:00:11.010 align:start position:0%
almost nothing and can do almost nothing
 

00:00:11.010 --> 00:00:12.890 align:start position:0%
almost nothing and can do almost nothing
for<00:00:11.160><c> ourselves</c><00:00:11.280><c> but</c><00:00:12.059><c> soon</c><00:00:12.360><c> we're</c><00:00:12.750><c> learning</c>

00:00:12.890 --> 00:00:12.900 align:start position:0%
for ourselves but soon we're learning
 

00:00:12.900 --> 00:00:15.589 align:start position:0%
for ourselves but soon we're learning
and<00:00:13.230><c> becoming</c><00:00:13.590><c> more</c><00:00:13.740><c> capable</c><00:00:14.190><c> every</c><00:00:14.460><c> day</c><00:00:14.610><c> but</c>

00:00:15.589 --> 00:00:15.599 align:start position:0%
and becoming more capable every day but
 

00:00:15.599 --> 00:00:17.090 align:start position:0%
and becoming more capable every day but
did<00:00:15.719><c> you</c><00:00:15.839><c> know</c><00:00:16.020><c> that</c><00:00:16.049><c> computers</c><00:00:16.680><c> can</c><00:00:16.859><c> do</c><00:00:16.980><c> the</c>

00:00:17.090 --> 00:00:17.100 align:start position:0%
did you know that computers can do the
 

00:00:17.100 --> 00:00:19.760 align:start position:0%
did you know that computers can do the
same<00:00:17.930><c> machine</c><00:00:18.930><c> learning</c><00:00:19.350><c> brings</c><00:00:19.500><c> together</c>

00:00:19.760 --> 00:00:19.770 align:start position:0%
same machine learning brings together
 

00:00:19.770 --> 00:00:21.560 align:start position:0%
same machine learning brings together
statistics<00:00:20.520><c> and</c><00:00:20.699><c> computer</c><00:00:21.090><c> science</c><00:00:21.420><c> to</c>

00:00:21.560 --> 00:00:21.570 align:start position:0%
statistics and computer science to
 

00:00:21.570 --> 00:00:23.330 align:start position:0%
statistics and computer science to
enable<00:00:22.050><c> computers</c><00:00:22.560><c> to</c><00:00:22.740><c> learn</c><00:00:22.890><c> how</c><00:00:23.039><c> to</c><00:00:23.070><c> do</c><00:00:23.310><c> a</c>

00:00:23.330 --> 00:00:23.340 align:start position:0%
enable computers to learn how to do a
 

00:00:23.340 --> 00:00:25.550 align:start position:0%
enable computers to learn how to do a
given<00:00:23.609><c> task</c><00:00:23.820><c> without</c><00:00:24.630><c> being</c><00:00:24.990><c> programmed</c><00:00:25.230><c> to</c>

00:00:25.550 --> 00:00:25.560 align:start position:0%
given task without being programmed to
 

00:00:25.560 --> 00:00:28.310 align:start position:0%
given task without being programmed to
do<00:00:25.680><c> so</c><00:00:25.910><c> just</c><00:00:26.910><c> as</c><00:00:27.060><c> your</c><00:00:27.210><c> brain</c><00:00:27.480><c> uses</c><00:00:27.930><c> experience</c>

00:00:28.310 --> 00:00:28.320 align:start position:0%
do so just as your brain uses experience
 

00:00:28.320 --> 00:00:31.240 align:start position:0%
do so just as your brain uses experience
to<00:00:28.740><c> improve</c><00:00:29.070><c> at</c><00:00:29.220><c> a</c><00:00:29.310><c> task</c><00:00:29.550><c> so</c><00:00:30.179><c> crew</c><00:00:30.390><c> computers</c>

00:00:31.240 --> 00:00:31.250 align:start position:0%
to improve at a task so crew computers
 

00:00:31.250 --> 00:00:33.740 align:start position:0%
to improve at a task so crew computers
say<00:00:32.250><c> you</c><00:00:32.309><c> need</c><00:00:32.669><c> a</c><00:00:32.700><c> computer</c><00:00:33.210><c> that</c><00:00:33.300><c> can</c><00:00:33.480><c> tell</c>

00:00:33.740 --> 00:00:33.750 align:start position:0%
say you need a computer that can tell
 

00:00:33.750 --> 00:00:35.030 align:start position:0%
say you need a computer that can tell
the<00:00:33.899><c> difference</c><00:00:34.230><c> between</c><00:00:34.380><c> a</c><00:00:34.590><c> picture</c><00:00:34.890><c> of</c><00:00:34.980><c> a</c>

00:00:35.030 --> 00:00:35.040 align:start position:0%
the difference between a picture of a
 

00:00:35.040 --> 00:00:37.700 align:start position:0%
the difference between a picture of a
dog<00:00:35.219><c> and</c><00:00:35.610><c> a</c><00:00:35.700><c> picture</c><00:00:35.910><c> of</c><00:00:36.030><c> a</c><00:00:36.180><c> cat</c><00:00:36.500><c> you</c><00:00:37.500><c> could</c>

00:00:37.700 --> 00:00:37.710 align:start position:0%
dog and a picture of a cat you could
 

00:00:37.710 --> 00:00:39.590 align:start position:0%
dog and a picture of a cat you could
begin<00:00:37.860><c> by</c><00:00:38.070><c> feeding</c><00:00:38.579><c> it</c><00:00:38.790><c> images</c><00:00:39.210><c> and</c><00:00:39.360><c> telling</c>

00:00:39.590 --> 00:00:39.600 align:start position:0%
begin by feeding it images and telling
 

00:00:39.600 --> 00:00:40.850 align:start position:0%
begin by feeding it images and telling
it<00:00:39.690><c> this</c><00:00:40.140><c> one's</c><00:00:40.379><c> a</c><00:00:40.469><c> dog</c>

00:00:40.850 --> 00:00:40.860 align:start position:0%
it this one's a dog
 

00:00:40.860 --> 00:00:43.670 align:start position:0%
it this one's a dog
that<00:00:41.219><c> one's</c><00:00:41.460><c> a</c><00:00:41.579><c> cat</c><00:00:41.790><c> a</c><00:00:42.110><c> computer</c><00:00:43.110><c> program</c><00:00:43.500><c> to</c>

00:00:43.670 --> 00:00:43.680 align:start position:0%
that one's a cat a computer program to
 

00:00:43.680 --> 00:00:45.529 align:start position:0%
that one's a cat a computer program to
learn<00:00:43.860><c> will</c><00:00:44.129><c> seek</c><00:00:44.430><c> statistical</c><00:00:45.180><c> patterns</c>

00:00:45.529 --> 00:00:45.539 align:start position:0%
learn will seek statistical patterns
 

00:00:45.539 --> 00:00:47.150 align:start position:0%
learn will seek statistical patterns
within<00:00:45.780><c> the</c><00:00:45.989><c> data</c><00:00:46.200><c> that</c><00:00:46.440><c> will</c><00:00:46.559><c> enable</c><00:00:46.829><c> it</c><00:00:47.100><c> to</c>

00:00:47.150 --> 00:00:47.160 align:start position:0%
within the data that will enable it to
 

00:00:47.160 --> 00:00:49.430 align:start position:0%
within the data that will enable it to
recognize<00:00:47.760><c> a</c><00:00:47.789><c> cat</c><00:00:48.149><c> or</c><00:00:48.360><c> a</c><00:00:48.420><c> dog</c><00:00:48.510><c> in</c><00:00:48.930><c> the</c><00:00:49.050><c> future</c>

00:00:49.430 --> 00:00:49.440 align:start position:0%
recognize a cat or a dog in the future
 

00:00:49.440 --> 00:00:51.740 align:start position:0%
recognize a cat or a dog in the future
it<00:00:50.370><c> might</c><00:00:50.520><c> figure</c><00:00:50.760><c> out</c><00:00:50.969><c> on</c><00:00:51.210><c> its</c><00:00:51.360><c> own</c>

00:00:51.740 --> 00:00:51.750 align:start position:0%
it might figure out on its own
 

00:00:51.750 --> 00:00:53.569 align:start position:0%
it might figure out on its own
that<00:00:52.110><c> cats</c><00:00:52.350><c> have</c><00:00:52.469><c> shorter</c><00:00:52.860><c> noses</c><00:00:53.190><c> and</c><00:00:53.309><c> that</c>

00:00:53.569 --> 00:00:53.579 align:start position:0%
that cats have shorter noses and that
 

00:00:53.579 --> 00:00:55.459 align:start position:0%
that cats have shorter noses and that
dogs<00:00:53.789><c> come</c><00:00:54.059><c> in</c><00:00:54.149><c> a</c><00:00:54.239><c> larger</c><00:00:54.629><c> variety</c><00:00:55.140><c> of</c><00:00:55.170><c> sizes</c>

00:00:55.459 --> 00:00:55.469 align:start position:0%
dogs come in a larger variety of sizes
 

00:00:55.469 --> 00:00:57.700 align:start position:0%
dogs come in a larger variety of sizes
and<00:00:55.829><c> then</c><00:00:56.730><c> represent</c><00:00:57.300><c> that</c><00:00:57.420><c> information</c>

00:00:57.700 --> 00:00:57.710 align:start position:0%
and then represent that information
 

00:00:57.710 --> 00:01:00.830 align:start position:0%
and then represent that information
numerically<00:00:58.710><c> organizing</c><00:00:59.670><c> it</c><00:00:59.789><c> in</c><00:00:59.879><c> space</c><00:01:00.210><c> but</c>

00:01:00.830 --> 00:01:00.840 align:start position:0%
numerically organizing it in space but
 

00:01:00.840 --> 00:01:03.139 align:start position:0%
numerically organizing it in space but
crucially<00:01:01.710><c> it's</c><00:01:02.010><c> the</c><00:01:02.190><c> computer</c><00:01:02.699><c> not</c><00:01:02.969><c> the</c>

00:01:03.139 --> 00:01:03.149 align:start position:0%
crucially it's the computer not the
 

00:01:03.149 --> 00:01:04.609 align:start position:0%
crucially it's the computer not the
programmer<00:01:03.629><c> that</c><00:01:03.719><c> identifies</c><00:01:04.379><c> those</c>

00:01:04.609 --> 00:01:04.619 align:start position:0%
programmer that identifies those
 

00:01:04.619 --> 00:01:06.620 align:start position:0%
programmer that identifies those
patterns<00:01:04.949><c> and</c><00:01:05.220><c> establishes</c><00:01:06.000><c> the</c><00:01:06.150><c> algorithm</c>

00:01:06.620 --> 00:01:06.630 align:start position:0%
patterns and establishes the algorithm
 

00:01:06.630 --> 00:01:08.960 align:start position:0%
patterns and establishes the algorithm
by<00:01:06.780><c> which</c><00:01:06.810><c> future</c><00:01:07.290><c> data</c><00:01:07.619><c> will</c><00:01:07.860><c> be</c><00:01:07.979><c> sorted</c><00:01:08.340><c> one</c>

00:01:08.960 --> 00:01:08.970 align:start position:0%
by which future data will be sorted one
 

00:01:08.970 --> 00:01:10.670 align:start position:0%
by which future data will be sorted one
example<00:01:09.479><c> of</c><00:01:09.540><c> a</c><00:01:09.600><c> simple</c><00:01:09.840><c> yet</c><00:01:10.170><c> highly</c><00:01:10.470><c> effective</c>

00:01:10.670 --> 00:01:10.680 align:start position:0%
example of a simple yet highly effective
 

00:01:10.680 --> 00:01:12.830 align:start position:0%
example of a simple yet highly effective
algorithm<00:01:11.280><c> is</c><00:01:11.670><c> to</c><00:01:11.729><c> find</c><00:01:12.150><c> the</c><00:01:12.270><c> optimal</c><00:01:12.630><c> line</c>

00:01:12.830 --> 00:01:12.840 align:start position:0%
algorithm is to find the optimal line
 

00:01:12.840 --> 00:01:15.020 align:start position:0%
algorithm is to find the optimal line
separating<00:01:13.500><c> cats</c><00:01:13.710><c> from</c><00:01:13.830><c> dogs</c><00:01:14.130><c> when</c><00:01:14.939><c> the</c>

00:01:15.020 --> 00:01:15.030 align:start position:0%
separating cats from dogs when the
 

00:01:15.030 --> 00:01:16.880 align:start position:0%
separating cats from dogs when the
computer<00:01:15.450><c> sees</c><00:01:15.659><c> a</c><00:01:15.689><c> new</c><00:01:15.930><c> picture</c><00:01:16.170><c> it</c><00:01:16.619><c> checks</c>

00:01:16.880 --> 00:01:16.890 align:start position:0%
computer sees a new picture it checks
 

00:01:16.890 --> 00:01:18.380 align:start position:0%
computer sees a new picture it checks
which<00:01:17.070><c> side</c><00:01:17.340><c> of</c><00:01:17.369><c> the</c><00:01:17.520><c> line</c><00:01:17.640><c> it</c><00:01:17.820><c> falls</c><00:01:18.030><c> on</c><00:01:18.180><c> and</c>

00:01:18.380 --> 00:01:18.390 align:start position:0%
which side of the line it falls on and
 

00:01:18.390 --> 00:01:22.520 align:start position:0%
which side of the line it falls on and
then<00:01:18.570><c> says</c><00:01:18.810><c> either</c><00:01:18.960><c> cat</c><00:01:19.530><c> or</c><00:01:19.920><c> dog</c><00:01:21.439><c> but</c><00:01:22.439><c> of</c>

00:01:22.520 --> 00:01:22.530 align:start position:0%
then says either cat or dog but of
 

00:01:22.530 --> 00:01:26.539 align:start position:0%
then says either cat or dog but of
course<00:01:22.770><c> there</c><00:01:22.979><c> can</c><00:01:23.009><c> be</c><00:01:23.130><c> mistakes</c><00:01:25.340><c> the</c><00:01:26.340><c> more</c>

00:01:26.539 --> 00:01:26.549 align:start position:0%
course there can be mistakes the more
 

00:01:26.549 --> 00:01:28.219 align:start position:0%
course there can be mistakes the more
data<00:01:26.759><c> the</c><00:01:26.970><c> computer</c><00:01:27.090><c> receives</c><00:01:27.689><c> the</c><00:01:28.080><c> more</c>

00:01:28.219 --> 00:01:28.229 align:start position:0%
data the computer receives the more
 

00:01:28.229 --> 00:01:29.780 align:start position:0%
data the computer receives the more
finely<00:01:28.740><c> tuned</c><00:01:29.009><c> is</c><00:01:29.159><c> how</c><00:01:29.310><c> the</c><00:01:29.369><c> rhythm</c><00:01:29.579><c> becomes</c>

00:01:29.780 --> 00:01:29.790 align:start position:0%
finely tuned is how the rhythm becomes
 

00:01:29.790 --> 00:01:31.850 align:start position:0%
finely tuned is how the rhythm becomes
and<00:01:30.390><c> the</c><00:01:30.479><c> more</c><00:01:30.630><c> accurate</c><00:01:31.110><c> it</c><00:01:31.229><c> can</c><00:01:31.290><c> be</c><00:01:31.530><c> in</c><00:01:31.710><c> its</c>

00:01:31.850 --> 00:01:31.860 align:start position:0%
and the more accurate it can be in its
 

00:01:31.860 --> 00:01:35.480 align:start position:0%
and the more accurate it can be in its
predictions<00:01:33.590><c> machine</c><00:01:34.590><c> learning</c><00:01:34.950><c> is</c><00:01:35.070><c> already</c>

00:01:35.480 --> 00:01:35.490 align:start position:0%
predictions machine learning is already
 

00:01:35.490 --> 00:01:37.789 align:start position:0%
predictions machine learning is already
widely<00:01:35.820><c> applied</c><00:01:36.299><c> the</c><00:01:37.020><c> technology</c><00:01:37.650><c> behind</c>

00:01:37.789 --> 00:01:37.799 align:start position:0%
widely applied the technology behind
 

00:01:37.799 --> 00:01:39.020 align:start position:0%
widely applied the technology behind
facial<00:01:38.310><c> recognition</c>

00:01:39.020 --> 00:01:39.030 align:start position:0%
facial recognition
 

00:01:39.030 --> 00:01:42.050 align:start position:0%
facial recognition
CLEC<00:01:39.570><c> scanned</c><00:01:39.869><c> speech</c><00:01:40.170><c> recognition</c><00:01:41.060><c> spam</c>

00:01:42.050 --> 00:01:42.060 align:start position:0%
CLEC scanned speech recognition spam
 

00:01:42.060 --> 00:01:44.870 align:start position:0%
CLEC scanned speech recognition spam
filters<00:01:42.509><c> on</c><00:01:42.630><c> your</c><00:01:42.659><c> Inbox</c><00:01:43.340><c> online</c><00:01:44.340><c> shopping</c><00:01:44.759><c> or</c>

00:01:44.870 --> 00:01:44.880 align:start position:0%
filters on your Inbox online shopping or
 

00:01:44.880 --> 00:01:48.289 align:start position:0%
filters on your Inbox online shopping or
viewing<00:01:45.119><c> recommendations</c><00:01:46.880><c> credit</c><00:01:47.880><c> card</c><00:01:48.090><c> for</c>

00:01:48.289 --> 00:01:48.299 align:start position:0%
viewing recommendations credit card for
 

00:01:48.299 --> 00:01:50.840 align:start position:0%
viewing recommendations credit card for
detection<00:01:48.689><c> and</c><00:01:48.960><c> so</c><00:01:49.409><c> much</c><00:01:49.799><c> more</c><00:01:50.100><c> at</c><00:01:50.729><c> the</c>

00:01:50.840 --> 00:01:50.850 align:start position:0%
detection and so much more at the
 

00:01:50.850 --> 00:01:52.819 align:start position:0%
detection and so much more at the
University<00:01:51.390><c> of</c><00:01:51.420><c> Oxford</c><00:01:51.570><c> machine</c><00:01:52.409><c> learning</c>

00:01:52.819 --> 00:01:52.829 align:start position:0%
University of Oxford machine learning
 

00:01:52.829 --> 00:01:54.950 align:start position:0%
University of Oxford machine learning
researchers<00:01:53.460><c> are</c><00:01:53.579><c> combining</c><00:01:54.119><c> statistics</c><00:01:54.750><c> and</c>

00:01:54.950 --> 00:01:54.960 align:start position:0%
researchers are combining statistics and
 

00:01:54.960 --> 00:01:56.539 align:start position:0%
researchers are combining statistics and
computer<00:01:55.409><c> science</c><00:01:55.740><c> to</c><00:01:55.829><c> build</c><00:01:56.100><c> algorithms</c>

00:01:56.539 --> 00:01:56.549 align:start position:0%
computer science to build algorithms
 

00:01:56.549 --> 00:01:58.340 align:start position:0%
computer science to build algorithms
that<00:01:57.030><c> can</c><00:01:57.180><c> solve</c><00:01:57.390><c> more</c><00:01:57.719><c> complex</c><00:01:58.200><c> problems</c>

00:01:58.340 --> 00:01:58.350 align:start position:0%
that can solve more complex problems
 

00:01:58.350 --> 00:02:01.010 align:start position:0%
that can solve more complex problems
more<00:01:59.100><c> efficiently</c><00:01:59.390><c> using</c><00:02:00.390><c> less</c><00:02:00.570><c> computing</c>

00:02:01.010 --> 00:02:01.020 align:start position:0%
more efficiently using less computing
 

00:02:01.020 --> 00:02:03.740 align:start position:0%
more efficiently using less computing
power<00:02:01.200><c> for</c><00:02:02.040><c> medical</c><00:02:02.399><c> diagnosis</c><00:02:03.119><c> to</c><00:02:03.390><c> social</c>

00:02:03.740 --> 00:02:03.750 align:start position:0%
power for medical diagnosis to social
 

00:02:03.750 --> 00:02:04.810 align:start position:0%
power for medical diagnosis to social
media

00:02:04.810 --> 00:02:04.820 align:start position:0%
media
 

00:02:04.820 --> 00:02:06.850 align:start position:0%
media
the<00:02:05.480><c> potential</c><00:02:05.990><c> of</c><00:02:06.050><c> machine</c><00:02:06.380><c> learning</c><00:02:06.560><c> to</c>

00:02:06.850 --> 00:02:06.860 align:start position:0%
the potential of machine learning to
 

00:02:06.860 --> 00:02:08.800 align:start position:0%
the potential of machine learning to
transform<00:02:07.460><c> our</c><00:02:07.729><c> world</c><00:02:07.760><c> is</c><00:02:08.300><c> truly</c>

00:02:08.800 --> 00:02:08.810 align:start position:0%
transform our world is truly
 

00:02:08.810 --> 00:02:12.850 align:start position:0%
transform our world is truly
mind-blowing<00:02:11.350><c> to</c><00:02:12.350><c> find</c><00:02:12.500><c> out</c><00:02:12.680><c> more</c><00:02:12.800><c> about</c>

00:02:12.850 --> 00:02:12.860 align:start position:0%
mind-blowing to find out more about
 

00:02:12.860 --> 00:02:15.010 align:start position:0%
mind-blowing to find out more about
machine<00:02:13.400><c> learning</c><00:02:13.430><c> visit</c><00:02:14.030><c> Oxford</c><00:02:14.420><c> sparks</c><00:02:14.750><c> dr.</c>

00:02:15.010 --> 00:02:15.020 align:start position:0%
machine learning visit Oxford sparks dr.
 

00:02:15.020 --> 00:02:17.320 align:start position:0%
machine learning visit Oxford sparks dr.
extra<00:02:15.350><c> AC</c><00:02:15.530><c> UK</c><00:02:16.100><c> follow</c><00:02:16.640><c> us</c><00:02:16.790><c> on</c><00:02:16.910><c> Twitter</c><00:02:17.120><c> and</c>

00:02:17.320 --> 00:02:17.330 align:start position:0%
extra AC UK follow us on Twitter and
 

00:02:17.330 --> 00:02:20.350 align:start position:0%
extra AC UK follow us on Twitter and
find<00:02:17.660><c> us</c><00:02:17.840><c> on</c><00:02:17.930><c> Facebook</c>