/[aleader]/aleader/doc/aleader-ref.texi
ViewVC logotype

Diff of /aleader/doc/aleader-ref.texi

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.4 by jpritikin, Sat Aug 16 10:46:27 2003 UTC revision 1.5 by jpritikin, Tue Sep 23 05:00:05 2003 UTC
# Line 200  constructive personality development. Line 200  constructive personality development.
200  @tab the situation, prior to any conceptual analysis or description  @tab the situation, prior to any conceptual analysis or description
201  @item abstract situation  @item abstract situation
202  @tab the simplified, objective, verbal description of a given actual situation  @tab the simplified, objective, verbal description of a given actual situation
203  @item actual emotion  @item actual affect
204  @tab the feeling, prior to any conceptual analysis or description  @tab the feeling, prior to any conceptual analysis or description
205  @item abstract emotion  @item affect term
206  @tab the simplified, objective, verbal description of a given actual emotion  @tab the simplified, objective, verbal description of a given actual affect
207  @item exemplar  @item exemplar
208  @tab the correct answers; the abstract representation of a film  @tab the correct answers; the abstract representation of a film
209  @end multitable  @end multitable
210    
211  An abstract situation is an evocative example for an actual emotion.  An  An abstract situation is an evocative example for an actual affect.  An
212  abstract emotion is a name assigned to a group of similar abstract  affect term is a name assigned to a group of similar abstract
213  situations.  situations.
214    
215  @node Audience, Bibliography, Terminology, Introduction  @node Audience, Bibliography, Terminology, Introduction
# Line 245  a full annotated bibliography is a long Line 245  a full annotated bibliography is a long
245  @chapter Classification  @chapter Classification
246    
247  We would like some way to evoke a repeatable sequence of atomic  We would like some way to evoke a repeatable sequence of atomic
248  emotions.  In general, film contains too much detail and complexity for  affects.  In general, film contains too much detail and complexity for
249  this purpose.  There are many problems:  this purpose.  There are many problems:
250    
251  @itemize @bullet  @itemize @bullet
252  @item A given personality in a given film clip may be depicting  @item A given personality in a given film clip may be depicting
253  several different emotions at once.  several different affects at once.
254    
255  @item The set of personalities who are simultaneously present in a film clip  @item The set of personalities who are simultaneously present in a film clip
256  need not be depicting similar emotions.  need not be depicting similar affects.
257    
258  @item The overall emotional tone depicted by a film clip can  @item The overall affective tone depicted by a film clip can
259  be distinct from what each of the individual personality in  be distinct from what each of the individual personality in
260  a film is depicting.  a film is depicting.
261  @end itemize  @end itemize
# Line 263  a film is depicting. Line 263  a film is depicting.
263  Without addressing these problems, an unambiguous analysis of film is  Without addressing these problems, an unambiguous analysis of film is
264  impossible.  The solution employed in Aleader is to prescribe rules  impossible.  The solution employed in Aleader is to prescribe rules
265  which skillfully narrow down on specific component of a scene.  We are  which skillfully narrow down on specific component of a scene.  We are
266  not interested in the whole complex mixture of emotions.  Rather, we are  not interested in the whole complex mixture of affects.  Rather, we are
267  trying to pick out one or two emotions as guided by carefully crafted  trying to pick out one or two affects as guided by carefully crafted
268  appraisal instructions.  appraisal instructions.
269    
270  We use two different tools to analyze a film:  We use two different tools to analyze a film:
271    
272  @enumerate  @enumerate
273  @item We use our innate sense of empathy to get a feel for the action  @item We use our innate mindreading sense to get a feel for the action
274  then pick an evocative phrase which best fits our feeling.  then pick an evocative phrase which best fits our feeling.
275    
276  @item The other method is described in this chapter.  @item The other method is described in this chapter.
# Line 518  the @code{Change Span} button.  If there Line 518  the @code{Change Span} button.  If there
518  situations (for example a conversation during a long span of @w{[+]  situations (for example a conversation during a long span of @w{[+]
519  admires [0]}) then the longer running spans should be ignored.  admires [0]}) then the longer running spans should be ignored.
520    
521  If we try to model all the layers of emotion ranging from one second spans  If we try to model all the layers of affect ranging from one second spans
522  to hour long spans then we invite too much uncertainty.  to hour long spans then we invite too much uncertainty.
523  When exactly do long span emotions start and end?  Exactly how many  When exactly do long span affects start and end?  Exactly how many
524  layers of emotion are presently active?  layers of affect are presently active?
525  There questions may not be answerable.  There questions may not be answerable.
526  Therefore, only consider situations of shortest span.  Therefore, only consider situations of shortest span.
527    
# Line 607  good judgment. Line 607  good judgment.
607  @end itemize  @end itemize
608    
609  @quotation  @quotation
610  Which factors @emph{actually} contribute to the emotional  Which factors @emph{actually} contribute to the affectual
611  texture of the film?  texture of the film?
612  @end quotation  @end quotation
613    
# Line 621  follows: Line 621  follows:
621    
622  @enumerate  @enumerate
623  @item  @item
624  Predict empathy: Map abstract situations to abstract emotions.  Predict affect: Map abstract situations to affect terms.
625    
626  @item  @item
627  Compare the abstract emotions with actual emotions.  Compare the affect terms with actual affects.
628  @end enumerate  @end enumerate
629    
630  If you feel the emotion ``nervousness'' but the calculated  If you feel the affect ``nervousness'' but the calculated
631  abstract emotion is ``drunk with accomplishment'' then something is  affect term is ``drunk with accomplishment'' then something is
632  wrong.  The calculated and actual emotions must be made to reconcile.  wrong.  The calculated and actual affects must be made to reconcile.
633  How does this work in detail?  How does this work in detail?
634    
635  Notice that step (1) of the reconciliation process requires an  Notice that step (1) of the reconciliation process requires a
636  empathy mapping.  We also would like to present an inverse-of-empathy  mindreading mapping.  We also would like to present an inverse-of-mindreading
637  mapping (situation assessment), otherwise we invite doubt about its  mapping (situation assessment), otherwise we invite doubt about its
638  existence.  A one-to-one mapping would be ideal.  If this mapping is  existence.  A one-to-one mapping would be ideal.  If this mapping is
639  merely one-to-many (or one-way) then the whole viability of the  merely one-to-many (or one-way) then the whole viability of the
# Line 642  merely one-to-many (or one-way) then the Line 642  merely one-to-many (or one-way) then the
642  @cartouche  @cartouche
643  @multitable @columnfractions .6 .4  @multitable @columnfractions .6 .4
644  @item  @item
645  Here is a typical situation/emotion entry.  Here is a typical situation/affect entry.
646  The part between @code{if} and @code{then} is the abstract situation pattern.  The part between @code{if} and @code{then} is the abstract situation pattern.
647  The word (or words) between the double quotes is the abstract emotion.  The word (or words) between the double quotes is the affect term.
648  The words describing the emotion can be of any type.  The words describing the affect can be of any type.
649  There is no artificial limitation.  There is no artificial limitation.
650  Any evocative phrase serves as an abstract emotion.  Any evocative phrase serves as an affect term.
651  @tab  @tab
652  @example  @example
653  if  if
# Line 673  deserves to be scrapped. Line 673  deserves to be scrapped.
673    
674  To search for counterexamples, use the Cross Reference screen.  To search for counterexamples, use the Cross Reference screen.
675  The technique is to look at situations and sort them by similarity.  Use  The technique is to look at situations and sort them by similarity.  Use
676  your innate sense of empathy to judge similarity.  Which situations  your innate mindreading sense to judge similarity.  Which situations
677  trigger similar emotions?  Which situations are not similar?  If you don't  trigger similar affects?  Which situations are not similar?  If you don't
678  get a clear sense of the emotion on first viewing then repeatedly view  get a clear sense of the affect on first viewing then repeatedly view
679  the span until you get it.  Counterexamples are of two kinds:  the span until you get it.  Counterexamples are of two kinds:
680    
681  @itemize @bullet  @itemize @bullet
682  @item  @item
683  The parameters of the situation are too general.  The parameters of the situation are too general.
684  Examples of a single situation type evoke more than one emotion.  Examples of a single situation type evoke more than one affect.
685    
686  For a given abstract situation, all matching film examples should evoke a  For a given abstract situation, all matching film examples should evoke a
687  similar emotion.  You can check by playing each example.  If a problem  similar affect.  You can check by playing each example.  If a problem
688  is found then try to solve it by adding a more specific pattern.  All of  is found then try to solve it by adding a more specific pattern.  All of
689  the large patterns were developed this way.  Emotions with a relatively  the large patterns were developed this way.  Affects with a relatively
690  large proportion of matches are good canidates for pattern refinement.  large proportion of matches are good canidates for pattern refinement.
691  There are some likely canidates in the default mapping.  These canidates  There are some likely canidates in the default mapping.  These canidates
692  are assigned multiple similar emotions separated by slashes.  are assigned multiple similar affects separated by slashes.
693    
694    
695  @item  @item
696  The parameters of situations are too specific.  The parameters of situations are too specific.
697  Examples of more than one situation type evoke a single emotion.  Examples of more than one situation type evoke a single affect.
698    
699  To detect this kind of problem, ignore situation types and look for abstract  To detect this kind of problem, ignore situation types and look for
700  emotions which are the same or evocate a similar feeling.  If two  affect terms which are the same or evocate a similar feeling.  If two
701  emotions are indistinguishable then there is a problem in the empathy  affects are indistinguishable then there is a problem in the mindreading
702  mapping.  mapping.
703  @end itemize  @end itemize
704    
705  Try to find a counterexample.  It's not easy, is it?  Gain confidence  Try to find a counterexample.  It's not easy, is it?  Gain confidence
706  that the empathy mapping is correct. Now we are ready for step (2).  that the mindreading mapping is correct. Now we are ready for step (2).
707    
708  Go through your abstract story.  Check whether the calculated emotions  Go through your abstract story.  Check whether the calculated affect terms
709  match the actual emotions.  If you find an emotion that doesn't match then the  match the actual affects.  If you find an affect that doesn't match then the
710  problem is with the empathy mapping or your abstract story.  Since you  problem is with the mindreading mapping or your abstract story.  Since you
711  already verified the empathy mapping, the fault can only be attributed  already verified the mindreading mapping, the fault can only be attributed
712  to your situation assessment ability.  to your situation assessment ability.
713    
714  @tex  @tex
# Line 750  parameters of an abstract situation rema Line 750  parameters of an abstract situation rema
750  @item Chaining  @item Chaining
751  Is this situation part of a conversation?  Is this situation part of a conversation?
752  @item Disagreement  @item Disagreement
753  Which factors @emph{actually} contribute to the emotional  Which factors @emph{actually} contribute to the affectual
754  texture of the film?  texture of the film?
755  @end table  @end table
756    
# Line 1019  Transcript translation is the obvious pr Line 1019  Transcript translation is the obvious pr
1019  translation, without dubbing the film, still demands the tedious process of  translation, without dubbing the film, still demands the tedious process of
1020  synchronizing the film with the translated transcript  synchronizing the film with the translated transcript
1021  (@pxref{Synchronization}) and then adjusting all the spans.  It's a lot  (@pxref{Synchronization}) and then adjusting all the spans.  It's a lot
1022  of manual work which i don't really know how to automate.  of manual work which we don't really know how to automate.
1023    
1024  @item  @item
1025  Within Aleader, there is a component which converts abstract situations  Within Aleader, there is a component which converts abstract situations
1026  into sentences.  No attempt is made to adapt the sentence according to  into sentences.  No attempt is made to adapt the sentence according to
1027  the gender of the participants.  No attempt is made to conjugate verb  the gender of the participants.  No attempt is made to conjugate verb
1028  tense.  It reads awkwardly.  i'm not sure whether other languages are  tense.  It reads awkward.  We have not investigated whether other
1029  more strict about such things.  languages are more strict about such things.
1030    
1031  @item  @item
1032  It is an open question how well the empathy mapping translates into  It is an open question how well the mindreading mapping translates into
1033  other languages.  i expect a core subset of emotions will translate into  other languages and cultures.
1034    We expect a core subset of affects will translate into
1035  all languages.  However, some languages may offer more names for  all languages.  However, some languages may offer more names for
1036  emotions than others languages.  i am curious to see what this direction  affects than others languages.  We are curious to see what this direction
1037  of inquiry will reveal.  of inquiry will reveal.
1038  @end itemize  @end itemize
1039    

Legend:
Removed from v.1.4  
changed lines
  Added in v.1.5

savannah-hackers-public@gnu.org
ViewVC Help
Powered by ViewVC 1.1.26