🎁Amazon Prime 📖Kindle Unlimited 🎧Audible Plus 🎵Amazon Music Unlimited 🌿iHerb 💰Binance
Video
Transcript
we tried to build robots that learn a
little bit like humans do by trial and
error what we’ve done is trained an
algorithm to solve the Rubik’s Cube
one-handed with a robotic captain which
is actually pretty hard even for a human
to do we don’t tell it how the hand is
to move the the cube in order to get
there the particular friction that’s on
the fingers how easy it is to turn the
faces on the cube what the gravity what
the weight of the cube is all of these
things it needs to learn by itself
the interesting thing is that kind of
standard techniques in robotics haven’t
been able to scale to that complexity
that we see in a robotic hand humans
have evolved to be able to manipulate
and operate our hands so there’s a huge
amount of learning that’s happened
through evolution to get us to this
point as a as a species and the robot
has to learn all of this from scratch
instead of trying to write very
dedicated algorithms to operate such a
hand we took a different approach where
we create thousands of different
simulated environments and learn to do
the task in all of those and hopefully
the robotic hand will be able to do it
in the real world as well this means
like thousands of years of experience
that is your network has had in
simulation every time the argument good
at the task we make the task harder
that’s really crucial because you need
exposure to really complicate
environments in order to eventually be
robust to the real world you put a
rubber glove on their hand and can still
carry out the task
this ability to generalize to new
environments feels like a very poor
piece of intelligence it really changes
the way we think about training of
general purpose robots
moving from thinking too much about the
actual arguments and start thinking
about how do we create complex enough
worlds where they can learn at some
point then it would be more down to the
imagination
what robots could actually accomplish
they hope is to build robots that can do
many different tasks to increase the
standard of living and give everybody a
better life
[Music]