Atomspace dynamics visualisation

[Juffin Hally]

As far as I'm aware, there're ways to visualize atoms in a static atomspace -- the AtomSpaceExplorer can be used for that. It's also possible to post/get atoms to/from ana atomspace and see the results in an Explorer as well.
But I couldn't find any correct & still working way of using the publisher/subscriber. So, if I create a new atom and link it to some other atom, I'd currently have to refresh the explorer page manually. And if I want to see some value change in real time (say, an attention value of some atom), I don't really know how to.

So, are there any examples of dynamics visualisation?

The best I could find was the workbench repo, which apparently was created 5 years ago and doesn't seem to be maintained.

[Ben Goertzel]

We had some dynamic visualizations of certain quantities... emotion
levels within OpenPsi, some parameters of ECAN etc. .. .but never a
dynamic visualization of Atomspace itself...

> --
> You received this message because you are subscribed to the Google Groups "opencog" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to opencog+u...@googlegroups.com.
> To post to this group, send email to ope...@googlegroups.com.
> Visit this group at https://groups.google.com/group/opencog.
> To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/5904ef06-7075-411f-85bc-f18cfa3e3569%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.



--
Ben Goertzel, PhD
http://goertzel.org

"Listen: This world is the lunatic's sphere, / Don't always agree
it's real. / Even with my feet upon it / And the postman knowing my
door / My address is somewhere else." -- Hafiz

[Juffin Hally]

Perhaps writing such a visualiser can be attempted? Is there any actual need for it (apart from me thinking it'd be helpful, I mean)?

Another point would be documenting the attempt; there might be someone working on it right now, but in secret, so far as I can tell from other public resources such as the wiki and the slack chat rooms.

On Monday, April 8, 2019 at 11:10:04 PM UTC+3, Ben Goertzel wrote:

We had some dynamic visualizations of certain quantities... emotion
levels within OpenPsi, some parameters of ECAN etc. .. .but never a
dynamic visualization of Atomspace itself...

On Mon, Apr 8, 2019 at 6:34 PM Juffin Hally <thirve....@gmail.com> wrote:
>
> As far as I'm aware, there're ways to visualize atoms in a static atomspace -- the AtomSpaceExplorer can be used for that. It's also possible to post/get atoms to/from ana atomspace and see the results in an Explorer as well.
> But I couldn't find any correct & still working way of using the publisher/subscriber. So, if I create a new atom and link it to some other atom, I'd currently have to refresh the explorer page manually. And if I want to see some value change in real time (say, an attention value of some atom), I don't really know how to.
>
> So, are there any examples of dynamics visualisation?
>
> The best I could find was the workbench repo, which apparently was created 5 years ago and doesn't seem to be maintained.
>
> --
> You received this message because you are subscribed to the Google Groups "opencog" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to ope...@googlegroups.com.

[Linas Vepstas]

On Wed, Apr 10, 2019 at 1:58 AM Juffin Hally <thirve....@gmail.com> wrote:

Perhaps writing such a visualiser can be attempted? Is there any actual need for it (apart from me thinking it'd be helpful, I mean)?

Another point would be documenting the attempt; there might be someone working on it right now, but in secret, so far as I can tell from other public resources such as the wiki and the slack chat rooms.

I don't think anyone is intentionally working "in secret", although there is a tendency to work quietly, and not talk very much.  It would be cool if people talked about their work more often.  Anyway, I'm 99% sure no one is working on this now.  You are welcome to try it. (create a github issue to track progress, and post to mailing list if your actually getting somewhere.)

... but!! A few comments about the design of such a thing.  The atomspace does currently have a callback-notification mechanism for when atoms are added/removed. Its .. not used much (or at all??). I kind of very much want to rip it out, because its in the critical performance path.  Currently, you can add-remove atoms at about 500K/second and anything that slows that down would be bad. The current notification system is in that direct path, and, misused, can quite easily slow down that path by factors of 10x or 100x  .. even by a factor of 2x even if you were super-careful and clean and elegant. I don't want to pay a 5% penalty, nevermind a 2x penalty.

So it would be cool if you (or a  mythical "someone"?) figured out how to monitor the atomspace from a different thread, (i.e. outside of the critical performance path) so that if users don't care about notifications, they would not have to pay any penalty.  Exactly how to make this fast & efficient, I don't know -- its a challenging, hard problem. I just want to avoid the problem of paying a performance penalty for a "gee whiz" feature.

--linas

On Monday, April 8, 2019 at 11:10:04 PM UTC+3, Ben Goertzel wrote:

We had some dynamic visualizations of certain quantities... emotion
levels within OpenPsi, some parameters of ECAN etc. .. .but never a
dynamic visualization of Atomspace itself...

On Mon, Apr 8, 2019 at 6:34 PM Juffin Hally <thirve....@gmail.com> wrote:
>
> As far as I'm aware, there're ways to visualize atoms in a static atomspace -- the AtomSpaceExplorer can be used for that. It's also possible to post/get atoms to/from ana atomspace and see the results in an Explorer as well.
> But I couldn't find any correct & still working way of using the publisher/subscriber. So, if I create a new atom and link it to some other atom, I'd currently have to refresh the explorer page manually. And if I want to see some value change in real time (say, an attention value of some atom), I don't really know how to.
>
> So, are there any examples of dynamics visualisation?
>
> The best I could find was the workbench repo, which apparently was created 5 years ago and doesn't seem to be maintained.
>
> --
> You received this message because you are subscribed to the Google Groups "opencog" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to ope...@googlegroups.com.
> To post to this group, send email to ope...@googlegroups.com.
> Visit this group at https://groups.google.com/group/opencog.
> To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/5904ef06-7075-411f-85bc-f18cfa3e3569%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.



--
Ben Goertzel, PhD
http://goertzel.org

"Listen: This world is the lunatic's sphere,  /  Don't always agree
it's real.  /  Even with my feet upon it / And the postman knowing my
door / My address is somewhere else." -- Hafiz

--
You received this message because you are subscribed to the Google Groups "opencog" group.

To unsubscribe from this group and stop receiving emails from it, send an email to opencog+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.
Visit this group at https://groups.google.com/group/opencog.

To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/715c7720-d569-4cc8-a3b1-b609dbc3dafb%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

cassette tapes - analog TV - film cameras - you

[Linas Vepstas]

Oh, and if you are only interested in attention values, that is "a lot easier". The attention bank already wraps all calls to change the attention values; and adding instrumentation there, to track what is going on, "make sense". Its .. easy, even, so, yes, please do.   

--linas