Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Elaborating object referents #2

Open
shiwalimohan opened this issue Mar 25, 2014 · 2 comments
Open

Elaborating object referents #2

shiwalimohan opened this issue Mar 25, 2014 · 2 comments
Assignees
@shiwalimohan
Labels
question

Comments

@shiwalimohan
Copy link
Contributor

I do not understand why the following was implemented -

(<int> ^stack.segments.segment <seg>
          ^referents.objects <objs>
         -^stack.segment <seg>)

Why are we not looking at the top-most segment to elaborate referent objects in focus?

line 95 in https://github.com/InstructoSoar/rosie/blob/master/agent/_agent/elaborations/elaborate-interaction.soar
@amininger
Copy link
Contributor

I think it has to do with the possibility of the top segment not being
fully processed and having objects that we don't want to index into because
we haven't finalized what they will be. For instance, if we have multiple
results and we haven't chosen the correct one yet. I'm not totally sure of
this though.
-Aaron

On Tue, Mar 25, 2014 at 3:42 PM, Shiwali Mohan [email protected]:

I do not understand why the following was implemented -
'''
( ^stack.segments.segment
^referents.objects
-^stack.segment )
'''

Why are we not looking at the top-most segment to elaborate referent
objects in focus?

line 95 in
https://github.com/InstructoSoar/rosie/blob/master/agent/_agent/elaborations/elaborate-interaction.soar

Reply to this email directly or view it on GitHubhttps://github.com//issues/2
.

@shiwalimohan
Copy link
Contributor Author

It looks like removing the negated test in this rule should not affect behavior. Removing the negated test will also make it consistent with the theory of attention - every object (or any other item) that is being talked about is considered to be in focus on the conversation. I will test the change to be sure nothing breaks.

PLindes added a commit that referenced this issue Jan 5, 2019
@PLindes
Align Lucia with Rosie #2
cb30ecb 
There is a commit with this same name in the rosie project.
On the Lucia side the handle for "heard" was changed from "heard" to "heard1" in both the grammar and smem.  Both comprehend-done and next-word were edited so that the same file works in both systems using some different apply rules.
On the Rosie side, several new concepts were added to lucia-extras, the changed xv-verbs files were copied in from Lucia, and Lucia_source.soar was changed to get the entire version A of comprehend from Lucia.  This last included the combined versions of comprehend-done and next-word.
Now Rosie runs the 209ordered.txt file in Lucia test mode with only on FAILURE, which is the expected one for "unkown."
How ever, there are two remaining issues which remain to be resolved:
On #86 the system stops for long periods on the interpret and do-action operators.
On #119 11 results with sequential numbers are printed out, sometimes.
Also, the code for sourcing from Lucia should be cleaned up somehow.
PLindes added a commit that referenced this issue May 25, 2021
@PLindes
Grounding: add TheKitchen construction
2cc55b0 
5/25/2021: A new TheKitchen construction was added to help getting "the kitchen" to ground either to the world or to smem. This adds five rules and works very nicely without any other changes. Part of the trace of running this in Rosie shows the preference based on lexicals, demonstrating semantic precision, and looks like this:

Received word #1: go
Matched a SimpleAction construction. C952
Received word #2: to
Received word #3: the
Received word #4: kitchen
Prefer >span: TheKitchen(2) > BareNoun(1).
Prefer >span: SpecNoun(2) > BareNoun(1).
Prefer >lexicals: TheKitchen(2) > SpecNoun(0).
Matched a TheKitchen construction. C971
Matched a SimplePrepPhrase construction. C975
Snipping C952 to get to C951 .
Attach PrepPhrase C975 to DriveVerb C951 .
Matched a DriveVerbToTarget construction. C977
Sentence #2: "go to the kitchen"

This makes "the kitchen" work very similarly to "the main office", whose trace looks like this:

Received word #1: go
Matched a SimpleAction construction. C1101
Received word #2: to
Received word #3: the
Received word #4: main
Received word #5: office
Prefer >span: TheMainOffice(3) > BareNoun(1).
Prefer >span: PropertyNoun(2) > BareNoun(1).
Prefer >span: SpecPropNoun(3) > BareNoun(1).
Prefer >lexicals: TheMainOffice(3) > SpecPropNoun(0).
Prefer >span: SpecPropNoun(3) > PropertyNoun(2).
Prefer >span: TheMainOffice(3) > PropertyNoun(2).
Matched a TheMainOffice construction. C1125
Matched a SimplePrepPhrase construction. C1129
Snipping C1101 to get to C1100 .
Attach PrepPhrase C1129 to DriveVerb C1100 .
Matched a DriveVerbToTarget construction. C1131
Sentence #2: "go to the main office"
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@shiwalimohan shiwalimohan

Labels
question
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@shiwalimohan @amininger