Tip:
Highlight text to annotate it
X
Well, as President has mentioned,
there are lots of different sources of intelligence.
Thus, we talked a lot about intelligence in the human brain.
Of course, people have been trying to make computers intelligent in some way.
I’m interested in when we get people and computers working together at scale.
People are - one thought, social things,
machines can drive trains, they’re good at other things.
Of course, sometimes we try to get them working together.
When we put them together, then typically,
we get somebody using computers to help them do their taxes or something.
It’s not very stunning. And we only have one of them.
I’m interested when we get people and computers working together
at what I call web scale.
What’s web scale? Web scale of things on the web.
There are around a trillion web pages out there.
There are billions of people.
So when you look at the web, actually,
we haven’t started it and we don’t really know much about it.
We can imagine, as this artist did, what it would be look like.
We can take some data about the web and try and plot this
in two or three dimensions, but it doesn’t really tell us a lot
although we should be doing more of this websites.
But this is a huge scale of which I’d like people to be working together
and working together with machines.
So it’s not just our intelligence but machine intelligence.
We’re not doing it yet.
What we are doing though is to a certain extent
getting people to work together, scale.
And what they do when you connect them on the web, well,
they join Facebook and twitter and they make things like Wikipedia.
This isn’t totally people working in a (inaudible)(01:36)
coz they still have this - the problems of the boundaries
between different social networking silos which they climb between,
but they’re working together quite large scale only if people will.
These are people working together
but they’re not using the computer as intelligence.
The computer is helping them.
It’s being the instrument by which they communicate as running the web browser
but it’s not using its intelligence.
So, conversely, what happens if we get lots of computers and we connect them?
Well, basically, nothing until you program them
and then we program them and they do things like search for it.
After (inaudible)(02:08) intelligence,
they try to work out the weather forecast and things like that.
And when they do that, they do in fact work in large scales.
Remember, you put link to open data out there
That’s very exciting and important to do but they’re not using people.
They’re not also using people.
So what would it be like if we actually try to get the massive humanity
and the massive -- the machines to be working together
connected by the Internet?
Connected.
It’s all about connection, isn’t it?
What’s important -- most intelligence, intelligence is all about connections.
It’s about finding connections.
I suppose that we had a system out there on the web,
we had connections shared between people and machines.
Suppose, when I put somebody in my Facebook group
when I said that somebody was in that photo,
then I was connected with the way machines work.
Suppose half of the idea in my head and half of the idea in yours
could both be somehow put on to the web and connected.
It’s all about connections.
This is technology we’re calling semantic web technology.
It involves representing technology -
representing the connections on the web just by links.
It’s just like the web very much.
Very simple technology, technology in which we use your eyes
but it’s where we involve the computers connected connecting in and processing
the same information as the people are connecting and, therefore,
making themselves part of a new system.
To make it happen we need to take those --
all programs which aren’t web scale and webize them,
make them work at web scale.
So we can take these where people --
where we’ve approximated to this,
where we’ve got people working at web scale
but not with people -- with machines involved,
they may involve the machines by adding some --
meaning the machines can figure.
We call that semantics.
We expose the web of data in the social networking system.
These are all things that we’re doing at MIT
to try and get towards this central situation
where we got everybody working together at scale.
Maybe we have machines working together
but no -- people are not involved.
We try to add into those machine systems an understanding of the social aspects
that you need to be able to work with people
and understanding of accountability, appropriate use of information, privacy.
We might try to make machines aware of those social aspects.
We call it policy awareness.
So that all those ways, our ways in which we’re doing research
to take the exciting place that we are now,
to take you something which -
where we think it’s gonna be much more exciting
when we produce an infrastructure where machines are contributing
their intelligence and people are contributing their intelligence
as part of one great big infrastructure which will be a huge infrastructure
for innovation in the future.
Thank you.