Well that settles it! They'll just have to hire us. Then the universe will be back in order XD!
71 karma · joined October 18, 2016
Ensuring the assets are "CAD-level accurate" has a huge impact on making this work in practice. So much so, that our collaborators Yash Narang [1] and Iretiayo Akinola [2] have spent a number of years on just this - we are standing on the shoulders of giants!
Additionally, understanding the hardware well (including things like motor and controller dynamics) is non-negotiable. We learned the hard way that there is, in fact, no free lunch. You really have to get the deep understanding of physics and engineering right.
[1] https://scholar.google.com/citations?user=M3NuG7AAAAAJ&hl=en [2] https://scholar.google.com/citations?user=e1zesfMAAAAJ&hl=en
This means no need to do the rollouts on real robots first! Please check out https://sgs-rl.github.io for some transfer examples (albeit a simpler setting then the multi-part NIST board we examine in AlphaNIST).
You are giving me some inspiration, though, that it might be good to try to explain the related techniques in a lower-dimensional setting to help teach the concept more effectively! brb
We can answer questions! But also, I want to highlight that the "AlphaNIST" name is an homage to the very well known work from David Silver's research group @ Deepmind circa 2016->2018. This was when we were coming up very early in our research careers and their work obviously had a big impact on us!
It has taken longer to make some of those ideas work for areas as challenging as robotics where there are physical constraints to deal with (and admittedly in this paper, we do not show the transfer to real hardware quite yet). But, we have some glimmer of evidence it is looking possible here (in another recent paper) that we cranked on for the last 6 months and will be featured at CoRL: https://x.com/octi_zhang/status/2108583315544682914
We are also all mega-lurkers here so we were honored to be featured on the front page!
There are a lot of areas where predictive models make sense in the robotics stack, but doing it with "video world models" as is trendy this year is likely a bet in the wrong direction according to the evidence we have been amassing in the last 6 months.
But, I'm wondering is there any major difference between just clicking the "..." next to the song in the default interface and selecting "Download->Audio"?
I think we need a more rigorous definition to go beyond tautology
Q* or any X* for that matter is extremely common for referring to the optimal function under certain assumptions. (usually cost / reward structure).
Specifically, Pablo Román's effort to make this viable for Spanish [1]. I started before there was a website or anything, just watching free youtube videos. Now I'm definitely B2 or C1. I subscribe partially just to show my support even though I mostly consume native-level media now.
Similar and helpful along he lines of CI are [2] and [3].
[0] https://en.wikipedia.org/wiki/Input_hypothesis [1] https://dreamingspanish.com [2] https://refold.la/ [3] https://www.alljapaneseallthetime.com/
If you happen to see this, please check my profile for a working email!
What I have found works best is to basically forget about putting strong priors on the structure of your notes. Just let everything be flat - or at most, create one level of hierarchy for a specific project.
When you have a new thought, use search. I have a command that pulls up the advanced search tool plugin and fuzzy searches my term. I then quickly scan (visually as cards) the notes I created already. If there is something related to my idea I open that note and continue that thread. If it's new, I create a new note.
This can be improved even more by using tags liberally, but I find that just having a powerful search and taking the time to "check" before creating a new note works very well for my style of note taking!
AI systems are quite good at perception in a traditional sense (e.g. determining a red ball is in the sensor receptive field), but once actual cognition about how the red ball relates to the world around it becomes important, current approaches are limited.
I would argue that the first wave of perception is solved much more completely than the ability to reason about complex interactions of objects grounded in the physical world.
Disclaimer: I am a robotics research scientist
Think about it for a moment. Can you define human intelligence? Have you ever gotten in a debate with someone about this? Is there a commonly accepted way to designate intelligence that isn't somewhat controversial?
How will we ever define AGI if we still haven't even defined RoHI (Regular ol' Human Intelligence) sufficiently?
Somehow, this has a perfect selection of games that formed a great portion of my childhood (some of which I was too young to play myself, but I watched my cousin or dad play).
Amazing to teleport into these worlds through the browser only 15 years later.
[1] https://en.wikipedia.org/w/index.php?title=Danbury_shakes&re...
I am earnestly trying to go all in on ActivityPub these days.