MIT Embodied Intelligence

Пікірлер

@jazzvids 2 ай бұрын

Thank you for this! presentation starts 19:33 :)

@the_engineer97 4 ай бұрын

This is so intriguing. I am so inspired

@LeoTX1 6 ай бұрын

Thanks!

@herbertarnold6372 6 ай бұрын

Promo-SM 😕

@LeoTX1 7 ай бұрын

It's a good presentation. Very useful for me! Thanks a lot!

@jeremydy3340 7 ай бұрын

Talk starts 32:57

@Shintuku 7 ай бұрын

Does this presentation correspond to some paper? It would be nice to have access to the slides/citations, very interesting stuff

@jeremydy3340 8 ай бұрын

Talk starts at 16:30

@hansbleuer3346 9 ай бұрын

Interesting explanation

@CandidDate 10 ай бұрын

I think a sense of humor in robotics would lead to clownish appeal.

@DhruvMetha 10 ай бұрын

Starts at 15:25

@aennmatyasbarra-hunyor5506 10 ай бұрын

Great one, thank you! I would like to be part of it. One day it will be possible.

@JosephHeck 11 ай бұрын

Content actually starts at 17:30, and the speaker's audio starts at 18:30

@araldjean-charles3924 Жыл бұрын

For the initial conditions that work, have anybody look at how much wiggle room you have. Is there an epsilon-neighborhood of the initial state you can safely start from, and how small is epsilon?

@cbasile22 Жыл бұрын

is there any formal course that covers multi agent RL? I find it confusing thus far. Thanks!

@AngeloKrs878 Жыл бұрын

1:07 subtitles "my experience with drugs couldn't be better"

@dbp_patel_1994 Жыл бұрын

😂

@bhaskartripathi 2 жыл бұрын

I was always confused by MA-MDP. You made it look very simplistic. Mathematical notations were very concise and research paper ready.

@keeperofthelight9681 3 жыл бұрын

how to do convolution lstm and other things on JAX more tutorials please sir Mathew Johnson

@kshitijshekhar1144 2 жыл бұрын

flax is a high level nn library built on top of Jax, check out its documentation. It's a very new library, built for flexibility. And you can make a mark in that by making PRs

@ImtithalSaeed 3 жыл бұрын

u and a confuse me

@georgemu7464 3 жыл бұрын

Very insightful

@devjaiswal1685 3 жыл бұрын

Thank you sir

@adamantinebipartite4732 3 жыл бұрын

Nazi.

@LB-fx1kx 3 жыл бұрын

Great work!

@iandanforth 3 жыл бұрын

Really enjoyed the presentation. The 'Puzzle' slide is problematic. All three have 'lots of wiring', the camera has smaller wires in a better package.

@syedshahid8316 Жыл бұрын

I live in karachi Pakistan I like your

@p.z.8355 3 жыл бұрын

How do you linearize the KG without getting into exponential complexity ?

@p.z.8355 3 жыл бұрын

How do you combine selfsupervised learning with declarative knowledge ?

@pakistanbtsarmy2625 3 жыл бұрын

👌

@ImtithalSaeed 3 жыл бұрын

Which book I can refer to

@f150bc 3 жыл бұрын

The diehold foundation is along with the suspecious observers pushing a 12 thousand year cycle of super nova and magnetic reversal which brings a catastrophic event please debate them on their theory's they have tens of thousands of people following them. I fear that the theory might be partially right. Find them on KZbin under those names.please look into this thanking you in advance Carl.

@fredxu9826 3 жыл бұрын

This is a great talk. Personally I haven't had the prerequisite for manifold learning, but the idea behind hybrid message passing is quite profound. Just wandering: if you have a Bayesian GNN where the prior encodes the linear assumption, would that be equivalent to the GNN + PGM model presented here? or is there a limit to the expressiveness of a Bayesian prior?