Re: [agi] Re: putting models in your robots head

yes, you can have a prediction system which learns over time, but it is also true that If it has an inbuilt model for the general reaction of ppl in its environment it needent learn it. :) -- Artificial General Intelligence List: AGI Permalink:

Re: [agi] Re: putting models in your robots head

learning is important if you want to make a robot that moves around people, things and interacts with people. you might want to command it to do something it was not programmed for. On Thu, Oct 31, 2019 at 4:20 AM wrote: > Just like to say one thing, learning isnt mandatory for a.i. to >

Re: [agi] Re: putting models in your robots head

Just like to say one thing,  learning isnt mandatory for a.i. to function,  if you have all the models it needs, then it doesnt have any more than what you give it to begin with,  but there still might be a lot of sampling it needs to do to use them. --

Re: [agi] Re: putting models in your robots head

All is needed Architecturally wise the layers themselves can be arbitrary but for most sequence processing people will shout "Transformer!" and call it a day Most processes do follow a sorta-autoregressive thing and sequence modeling is a good way to look at the problem, so I get the jump to

Re: [agi] Re: putting models in your robots head

Learning is important question and I do not want to take away from it, but these can be treated independently -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M683fd2c37cb525d42124bb21 Delivery

Re: [agi] Re: putting models in your robots head

https://ai.stackexchange.com/questions/10623/what-is-self-supervised-learning-in-machine-learning Self supervised learning is also when you talk to yourself, you squeeze out and mine gold from what you know, so it's supervised. but not by us lol. Itself! Unsupervised! Anyway, it's learning, ok?

Re: [agi] Re: putting models in your robots head

self supervised learning is another one. is self supervised learning better than one shot imitation learning? what about self supervised imitation learning? is there a such thing? On Wed, Oct 30, 2019 at 7:04 PM wrote: > The issue for me is different generalized classes of tasks operate on >

Re: [agi] Re: putting models in your robots head

The issue for me is different generalized classes of tasks operate on different sets of symbols, and we often use AI that are trained to operate on a single set of symbols, often the symbols are dependent on the AI as well, as with GPT-2 token embedding and position embedding, or any

Re: [agi] Re: putting models in your robots head

Transformers make cool hive minds, narrow AI obviously -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M563c158dd2da494a07f0aa66 Delivery options: https://agi.topicbox.com/groups/agi/subscription

Re: [agi] Re: putting models in your robots head

 Me as the worlds top AGI scientist, GPT-2  is not AGI. It is still narrow AI.  AGI is made up of many  narrow AI's to work in the same body. But some of these narrow in the bunch generate other narrow AI's.  In a completer unsupervised manner.   I do not see these special AI's generating and

Re: [agi] Re: putting models in your robots head

unsupervidsed  one-shot  imitation-learning  inter-agent  self-play Did yous ever think GPT-2 could self-play itself through imitation? Imitation keeps the topic same but also learns knowledge quickly by passing it down, which requires communicating with other related agents.

Re: [agi] Re: putting models in your robots head

  I see unsupervised learning is still very primitive, in the world of machine learning. -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-Mb05f719a17457fbb2fafc0e6 Delivery options:

Re: [agi] Re: putting models in your robots head

here is link with a definition for one shot learning. https://paperswithcode.com/task/one-shot-learning one shot imitation learning. Our approach is to combine meta-*learning* with *imitation learning* to enable *one*-*shot imitation learning*. The core idea is that provided *a single*

Re: [agi] Re: putting models in your robots head

please explain? -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M20982232dbbd1a96d3535b44 Delivery options: https://agi.topicbox.com/groups/agi/subscription

Re: [agi] Re: putting models in your robots head

you could look up one shot learning. it is interesting. you could look up the definition of one shot imitation learning. On Wed, Oct 30, 2019 at 8:49 AM wrote: > Im not sure what that stuffs about - but if you can imitate something, you > can guess what itll do next. > *Artificial General

Re: [agi] Re: putting models in your robots head

Im not sure what that stuffs about - but if you can imitate something, you can guess what itll do next. -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M698c3eae44877dfac6491f1e Delivery

Re: [agi] Re: putting models in your robots head

that is cool Doddy. 3d computer vision is the way for me too, because it removes the single perspective from the detection.   The first computer vision was 2d, and it still works, but 3d is the cheeky alternative that should work just as well with less stuffing around with neural network mumbo

Re: [agi] Re: putting models in your robots head

facebook is doing it too. here is a link. https://venturebeat.com/2019/10/29/facebook-highlights-ai-that-converts-2d-objects-into-3d-shapes/ On Tue, Oct 29, 2019 at 2:44 PM doddy wrote: > you posted the video on youtube today. > > On Tue, Oct 29, 2019 at 12:42 PM wrote: > >> had that music

Re: [agi] Re: putting models in your robots head

you posted the video on youtube today. On Tue, Oct 29, 2019 at 12:42 PM wrote: > had that music blasting and was creepy > *Artificial General Intelligence List * > / AGI / see discussions + > participants

Re: [agi] Re: putting models in your robots head

had that music blasting and was creepy -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-Md7ea347282553305af6de7f9 Delivery options: https://agi.topicbox.com/groups/agi/subscription

Re: [agi] Re: putting models in your robots head

nothing happened? -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M10aec0635d346aee185f96fe Delivery options: https://agi.topicbox.com/groups/agi/subscription

Re: [agi] Re: putting models in your robots head

On Tuesday, October 29, 2019, at 8:47 AM, doddy wrote: > take a look at this video. > https://www.stereolabs.com/blog/zed-powers-olympic-ski-robots/ I just had a crazy experience while surfing the internet: https://www.youtube.com/watch?v=FYII_rYpkSc=youtu.be

Re: [agi] Re: putting models in your robots head

yeah you maybe right. On Tue, Oct 29, 2019 at 8:55 AM wrote: > computer vision totally wreaks havoc in the video editing room. there > should be more of this stuff out there. > *Artificial General Intelligence List * > / AGI / see discussions

Re: [agi] Re: putting models in your robots head

computer vision totally wreaks havoc in the video editing room.  there should be more of this stuff out there. -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M249bfff923bcda1aa6b8eb80 Delivery

Re: [agi] Re: putting models in your robots head

have you seen this.they produced this too https://www.voxel51.com/ On Tue, Oct 29, 2019 at 8:08 AM wrote: > Its not fake, its a good go, but like you said the design isnt that > impressive or ground breaking. > Boston Dynamics are nearly there with the parkour atlas video, that... is >

Re: [agi] Re: putting models in your robots head

Its not fake,   its a good go,  but like you said the design isnt that impressive or ground breaking. Boston Dynamics are nearly there with the parkour atlas video,  that... is decent, and a little disturbing. -- Artificial General Intelligence List: AGI

Re: [agi] Re: putting models in your robots head

you are saying that is fake. it might be because that is the way the ski robots are designed. On Tue, Oct 29, 2019 at 7:53 AM wrote: > Thats actually not that impressive, If I dont see the knees pumping its > not the real thing. > *Artificial General Intelligence List

Re: [agi] Re: putting models in your robots head

Thats actually not that impressive,   If I dont see the knees pumping its not the real thing. -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M989652952aa914a6e27b5808 Delivery options:

Re: [agi] Re: putting models in your robots head

take a look at this video. https://www.stereolabs.com/blog/zed-powers-olympic-ski-robots/ On Tue, Oct 29, 2019 at 7:38 AM wrote: > yeh thatll give you a head start, but if u want to put the work in one > web cam can do it by itself. > 3d cameras make compositing so much easier, its almost a

Re: [agi] Re: putting models in your robots head

yeh thatll give you a head start,   but if u want to put the work in one web cam can do it by itself. 3d cameras make compositing so much easier, its almost a crime to have one! :) -- Artificial General Intelligence List: AGI Permalink:

Re: [agi] Re: putting models in your robots head

take a look at this depth camera. depth sensing cameras can be used for in robots. here is a link. https://www.stereolabs.com/ zed some of these have chemicals that cause cancer. look for one does not. On Mon, Oct 28, 2019 at 7:15 PM Alan Grimes via AGI wrote: > doddy wrote: > > take a look at

Re: [agi] Re: putting models in your robots head

doddy wrote: > take a look at this. > https://gigazine.net/gsc_news/en/20181206-nvidia-ai-rendered-virtual-world This is getting ridiculous! We are so absurdly close to AGI it's insane! I mean even I could probably hack that into general intelligence and it wouldn't take me very long either...

Re: [agi] Re: putting models in your robots head

oops, i dont know how that happened sorry... -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M09d9ec0dbb141fbb1821f81a Delivery options: https://agi.topicbox.com/groups/agi/subscription

Re: [agi] Re: putting models in your robots head

yes,  and is the agent in the world, or is the world in the agent. :) -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M365c3adf0b7f8d93264ca56d Delivery options:

Re: [agi] Re: putting models in your robots head

yes,  and is the agent in the world, or is the world in the agent. :) -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-Mc64fffcfa406d835bd19ddf9 Delivery options:

Re: [agi] Re: putting models in your robots head

yes,  and is the agent in the world, or is the world in the agent. :) -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-Meec9f826adac5dcd9260e8b6 Delivery options:

Re: [agi] Re: putting models in your robots head

yes,  and is the agent in the world, or is the world in the agent. :) -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M0a1afa8bea94c41f20efb4e0 Delivery options:

Re: [agi] Re: putting models in your robots head

yes,  and is the agent in the world, or is the world in the agent. :) -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-Mf97f87fd2e1be6bc596ecffa Delivery options:

Re: [agi] Re: putting models in your robots head

yes,  and is the agent in the world, or is the world in the agent. :) -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M7e35880400896bd97e7902e9 Delivery options:

Re: [agi] Re: putting models in your robots head

yes,  and is the agent in the world, or is the world in the agent. :) -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M26cdeb98c3b34b4124278901 Delivery options:

Re: [agi] Re: putting models in your robots head

yes,  and is the agent in the world, or is the world in the agent. :) -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M49c024e39e16e1592cc0edf1 Delivery options:

Re: [agi] Re: putting models in your robots head

yes,  and is the agent in the world, or is the world in the agent. :) -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M114fa4f359495e186ee1d676 Delivery options:

Re: [agi] Re: putting models in your robots head

yes,  and is the agent in the world, or is the world in the agent. :) -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M07797ce8086000c6647c2ac6 Delivery options:

Re: [agi] Re: putting models in your robots head

yes,  and is the agent in the world, or is the world in the agent. :) -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-M59997f7b59986e8d1ec2c3ee Delivery options:

Re: [agi] Re: putting models in your robots head

i think you mean the agent representing the robot. On Mon, Oct 28, 2019 at 8:10 AM wrote: > that would be way cool if u ran your bot inside that.if you have a > fake virtual environment, the computer has full power over it, why > search cant do that in real life, is because theres hidden

Re: [agi] Re: putting models in your robots head

that would be way cool if u ran your bot inside that.    if you have a fake virtual environment,   the computer has full power over it,  why search cant do that in real life, is because theres hidden variables. And... of course...  it takes too much computational resources. :)

Re: [agi] Re: putting models in your robots head

take a look at this. https://gigazine.net/gsc_news/en/20181206-nvidia-ai-rendered-virtual-world On Mon, Oct 28, 2019 at 7:14 AM wrote: > Definitely, that u could combine it with some self learning aparatus that > you could wait to develop over time, instead of giving it everything. But >

Re: [agi] Re: putting models in your robots head

Definitely, that u could combine it with some self learning aparatus that you could wait to develop over time,  instead of giving it everything.  But thats all I plan on doing, putting the lot in, I think its more trustworthy if I dont let it work things out,  more secure.

Re: [agi] Re: putting models in your robots head

50, 100 whatever. Build and retain unrecognized model graphs in memory... robot encounters revolving door, model not pre-packed so SOM graph stored for future encounters. Robot stuck in revolving door signals for assist avoids door next encounter. Just a fancy FSM. It helps to architect within

Re: [agi] Re: putting models in your robots head

yes u get me. SOM could work, definitely id suggest unsupervised learning for its recognition, but the actual formulas and protocols is fully supervised. 50 states may accomplish quite a bit, but im only guessing yet, coding the modules is like high-level english symbolic commandments crossed

Re: [agi] Re: putting models in your robots head

Accomplishing that goal is much less daunting yet can lead to generalization and provides valuable reward feedback to the engineers :) Pre-pack like 50 models then SOM map from eye to model as selector? -- Artificial General Intelligence List: AGI

Re: [agi] Re: putting models in your robots head

Yes, you need to detect the states off its eye. If it detects a door in its way,  then it follows the door opening protocol,  if it meets stairs it follows the stairs protocol, if it is playing soccer, it follows the soccer protocol,  if its getting in and out of its store truck, etcetera.

Re: [agi] Re: putting models in your robots head

Yes I agree with that also, and doing that you still need some basic Model Selection Criterion though it could be very simple. -- Artificial General Intelligence List: AGI Permalink:

Re: [agi] Re: putting models in your robots head

If it were AGI it wouldnt just be a model, it would be a model that builds models. But im not interested in that,  I think if you provide the models for the robots activity inside a specific environment, then you can get really far already,  by providing the robot what it knows, instead of it

Re: [agi] Re: putting models in your robots head

Hand picked models are fine but with general intelligence you more need models that can model multiple things. Your chess modeler might model a subset of all games for example chess, checkers, bridge, etc... Or that modeler might produce model instances one of which would be chess. Or say a

Re: [agi] Re: putting models in your robots head

a cellular automata is comprised of so many rules,  and these are the laws of the cellular automata, which is not this physics, its artificial. If the model is complete or not, its up to you to put it in. -- Artificial General Intelligence List: AGI

Re: [agi] Re: putting models in your robots head

For that matter if you can prove that the universe can be modeled from a few bits of cellular automata then all models are complete and exact? -- Artificial General Intelligence List: AGI Permalink:

Re: [agi] Re: putting models in your robots head

makes sense.   but if u have all the chess rules, then you have the model in complete, and its not approximate, its exact. -- Artificial General Intelligence List: AGI Permalink:

Re: [agi] Re: putting models in your robots head

If you model the unknown it is not exact except for what you know. So you have to model check on what you know or model check your prediction... if that makes sense... who knows... checking... -- Artificial General Intelligence List: AGI Permalink:

Re: [agi] Re: putting models in your robots head

what about never ending procrastination!!!  that gets the company forward!! -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T8fe5317c3cebf70b-Me49f060c41d8a6f6ab0f020d Delivery options:

Re: [agi] Re: putting models in your robots head

All models are approximations and there is compute and time distance from the modeled. And then the model checker is important, I pursue dynamic models and dynamic checking: https://en.wikipedia.org/wiki/Model_checking -- Artificial General Intelligence

Re: [agi] Re: putting models in your robots head

I think you have to sample to do a.i.,   because whats the point if youve got a model but you arent sampling from it?  A physics engine isnt any good until you put enough through it.  thats the computational expense. -- Artificial General Intelligence

Re: [agi] Re: putting models in your robots head

yes, just think how sparse the letters are from each other, and theres only 26, nice and approximate, learns instantaniously. -- Artificial General Intelligence List: AGI Permalink:

Re: [agi] Re: putting models in your robots head

You need to have accurate ways temporally to disregard accuracy. Muilt-modelling based on computational resource availability choosing between quick and accurate. -- Artificial General Intelligence List: AGI Permalink:

Re: [agi] Re: putting models in your robots head

Keep a few gems but the rest are security rubbish,  inventors getting owny about what they do, delivers very poor explanations. -- Artificial General Intelligence List: AGI Permalink:

Re: [agi] Re: putting models in your robots head

Papers cause my brain instant pain. So long and boring On Sun, 27 Oct 2019 at 14:22, Brett N Martensen wrote: > You should read this - Building Machines That Learn and Think Like > Peopleby Lake, B. M. et al > http://web.stanford.edu/class/psych209/Readings/LakeEtAlBBS.pdf > pages 16-19

[agi] Re: putting models in your robots head

You should read this -   Building Machines That Learn and Think Like People    by Lake, B. M. et al http://web.stanford.edu/class/psych209/Readings/LakeEtAlBBS.pdf pages 16-19  Section 4.1.1 Intuitive physics -- Artificial General Intelligence List: AGI