Re: neural/symbolic integration workshop

Posted by Frank Wimberly-2 on
URL: http://friam.383.s1.nabble.com/neural-symbolic-integration-workshop-tp7600618p7600823.html

My erstwhile colleagues in causal reasoning developed a taxonomy of graphs which would arise during the process of executing the algorithms we developed for learning a graphical causal model from a dataset.  These involved directed, undirected, and bidirectional edges.  I don't know whether any of this would be useful in other applications.  One of the most important categories was a PAG or partial ancestral graph.

On Wed, Feb 17, 2021, 12:15 PM Stephen Guerin <[hidden email]> wrote:

Thanks for forwarding, Jon. I did a 5-minute skim and would agree with Glen. I would add that a bidirectional path tracer would reverse the meaning of the semantic edge. eg, the next sentence after Glen's quote:

Although we have directed edges in the schema graph, we traverse it
in an undirected manner: from any vertex v, we visit its neighbors
from both incoming and outgoing edges. R3 shows an example
path containing an inherited edge,
    Actor ---worksAt −−→ Organization

The inverse would be something like
             Actor <---Employs --- Organization

This is equivalent to changing the lane direction in a traffic routing algorithm when flood-filling backwards during the bidirectional path tracing.

-Stephen

On Wed, Feb 17, 2021 at 10:47 AM uǝlƃ ↙↙↙ <[hidden email]> wrote:

I didn't see that presentation. My attendance has been marred by a lack of work-life balance. But there is a tantalizing bidirectionality statement in the paper: "Although we have directed edges in the schema graph, we traverse it in an undirected manner: from any vertex v, we visit its neighbors from both incoming and outgoing edges." But maybe the interesting duality lies not in the ER vs DR schema but in construction-pruning, which might map better to the Feynman integral.

On 2/16/21 11:08 AM, jon zingale wrote:
> I missed the first day, and while so far the second hasn't met my
> expectations, it was cool to catch Yang-Chen's presentation on *Ontological
> Pathfinding*[1, 2]. StephenG, assuming you are out there, I am curious about
> your thoughts. How might your expansions on bidirectional path-tracing
> apply?
>
> [1] https://dl.acm.org/doi/10.1145/2882903.2882954
> [2] https://github.com/yang-chen/Ontological-Pathfinding


--
↙↙↙ uǝlƃ

- .... . -..-. . -. -.. -..-. .. ... -..-. .... . .-. .
FRIAM Applied Complexity Group listserv
Zoom Fridays 9:30a-12p Mtn GMT-6  bit.ly/virtualfriam
un/subscribe http://redfish.com/mailman/listinfo/friam_redfish.com
FRIAM-COMIC http://friam-comic.blogspot.com/
archives: http://friam.471366.n2.nabble.com/

- .... . -..-. . -. -.. -..-. .. ... -..-. .... . .-. .
FRIAM Applied Complexity Group listserv
Zoom Fridays 9:30a-12p Mtn GMT-6  bit.ly/virtualfriam
un/subscribe http://redfish.com/mailman/listinfo/friam_redfish.com
FRIAM-COMIC http://friam-comic.blogspot.com/
archives: http://friam.471366.n2.nabble.com/

- .... . -..-. . -. -.. -..-. .. ... -..-. .... . .-. .
FRIAM Applied Complexity Group listserv
Zoom Fridays 9:30a-12p Mtn GMT-6  bit.ly/virtualfriam
un/subscribe http://redfish.com/mailman/listinfo/friam_redfish.com
FRIAM-COMIC http://friam-comic.blogspot.com/
archives: http://friam.471366.n2.nabble.com/