Natural Language Processing and emotions

[Nicola Alexander Schlup]

Hi all

I would like to know if there are any possibilities to determine
emotions for given sentences with Computational Linguistics/Natural
Language Processing. Is there any known solution for this pattern? It
should not just be an AIML file where every sentence got a static
emotion code. I need such a solution for a research paper where I want
to let a chatbot add emotion code to given sentences. The output could
be used for avaters in 3D tools or any other visualization tool. Thanks
a lot.

Kind regards,
Nicola Alexander Schlup
_______________________________________________
This is the alicebot-general mailing list
Reply to alicebot-general@...
Unsubscribe and change preferences at http://list.alicebot.org/mailman/listinfo/alicebot-general
Learn netiquette at http://www.dtcc.edu/cs/rfc1855.html
Learn to read at http://www.literacy.org/

[Chris Lofting]

Hi Nicola,

> -----Original Message-----
> From: alicebot-general-bounces@... [mailto:alicebot-general-
> bounces@...] On Behalf Of Nicola Alexander Schlup
> Sent: Thursday, 25 January 2007 3:15 AM
> To: alicebot-general@...
> Subject: [alicebot-general] Natural Language Processing and emotions
>
> Hi all
>
> I would like to know if there are any possibilities to determine
> emotions for given sentences with Computational Linguistics/Natural
> Language Processing. Is there any known solution for this pattern? It
> should not just be an AIML file where every sentence got a static
> emotion code.

Firstly you need the categories for primary emotions:
 (http://members.iimetro.com.au/~lofting/emote.html )

Then you need the categories for secondary emotions - i.e. those dependent
on the presence of a sense of SELF (these emotions develop with that sense
over the first 24 months of life)

Then you need the concepts (rules of engagement) of emotional
stimulus/response OR you can build such using a random generator of
emotional colourings and develop 'best fit' rules from there OR do what
infants to - return the emotion presented as stimulus.

In basic interactions, especially those requiring rapid assessments, the
instincts/habits try to identify sameness within a difference. Education
then refines these through exposure to differences; the less education, the
less understanding of differences, the more expression that is based on
stereotyping etc etc.

Emotional expression is strongly associated with magnitudes and the
symmetric. This can be a problem with logic in that emotions can allow for
such illogical categories such as 'this man is evil' -> all men are evil ->
the label 'men' and 'evil' are interchangeable (and so the symmetry focus).

It is the ASYMMETRIC that allows us to 'calm the waters' through clear
identification of local context differences and so the particular. You can
see this sort of thing at work in such experiments as the "Emotional I
Ching" - http://members.iimetro.com.au/~lofting/myweb/EmotionalIC.html

In this work we can identify consciousness/reason censoring emotional
expressions due to the expression being inappropriate (social rules) or
unfounded in that consciousness picks up details that emotion cannot
not/does not detect.

I would suggest the references in the above page on emotions (emote.html)
and in particular Plutchik's work.

For the development of SELF-dependent emotions look up material on the
development of SELF and such emotions as embarrassment and issues of guilt
(personal) vs shame (social) etc.

Since emotions are more often associated with the unconscious so come issues
of dealing with refined expressions through implicit forms - images etc in
that communication with the unconscious is by implicit means and moves us
into consideration of innuendo etc.

The Emotional IC uses the images of the I Ching that reflect the emotions
assessment of a situation in that the isomorphism between emotion and the I
Ching is in the method used to derive the basic categories -
self-referencing. Thus self-referencing of fight/flight and yang/yin will
establish generic categories across both specialist perspectives (working
off the concept that all meaning is determined by the method used to derive
it - thus differences are in the LABELS rather than the underlying
categories - the universals presented in
http://members.iimetro.com.au/~lofting/myweb/introIDM.html )


Chris
-----------------------------
generic categories of meaning:
------------------------------
Objects bias (differentiating):
BLEND - wholeness, whole numbers
BOUND - partness, rational numbers
Relationships bias (integrating):
BOND - share space, irrational numbers
BIND - share time, imaginary numbers

>From these come composites as reals, complex, quaternions, octonions. All
else follows....
http://members.iimetro.com.au/~lofting/myweb/introIDM.html

_______________________________________________
This is the alicebot-general mailing list
Reply to alicebot-general@...
Unsubscribe and change preferences at http://list.alicebot.org/mailman/listinfo/alicebot-general
Learn netiquette at http://www.dtcc.edu/cs/rfc1855.html
Learn to read at http://www.literacy.org/

[Chris Lofting]

>
> Firstly you need the categories for primary emotions:
>  (http://members.iimetro.com.au/~lofting/emote.html )
>


That should be http://members.iimetro.com.au/~lofting/myweb/emote.html
_______________________________________________
This is the alicebot-general mailing list
Reply to alicebot-general@...
Unsubscribe and change preferences at http://list.alicebot.org/mailman/listinfo/alicebot-general
Learn netiquette at http://www.dtcc.edu/cs/rfc1855.html
Learn to read at http://www.literacy.org/

[Chris Lofting]

You may also be interested in this page:

http://www.conknet.com/~mmagnus/

>From a focus on phonetics etc it becomes possible to pick up implicit
meanings with emotional content that can be responded to.

Chris
-----------------------------
generic categories of meaning:
------------------------------
Objects bias (differentiating):
BLEND - wholeness, whole numbers
BOUND - partness, rational numbers
Relationships bias (integrating):
BOND - share space, irrational numbers
BIND - share time, imaginary numbers

>From these come composites as reals, complex, quaternions, octonions. All
else follows....
http://members.iimetro.com.au/~lofting/myweb/introIDM.html
 
_______________________________________________
This is the alicebot-general mailing list
Reply to alicebot-general@...
Unsubscribe and change preferences at http://list.alicebot.org/mailman/listinfo/alicebot-general
Learn netiquette at http://www.dtcc.edu/cs/rfc1855.html
Learn to read at http://www.literacy.org/

[XGargoyle]

Hi Nicola,

You can use as a base the excellent thesis by Siska Fitrianie back in August 2002. The document discuss the implemmentation of exactly what you're interested in.

The pdf is available here:
http://www.kbs.twi.tudelft.nl/Publications/MSc/2002-Fitriani-MSc.html

I hope this can be helpful.

Regards,
Adam "XGargoyle" Codony

Nicola Alexander Schlup wrote:

Hi all

I would like to know if there are any possibilities to determine
emotions for given sentences with Computational Linguistics/Natural
Language Processing.

[Gary Dubuque]

Hi Nicola,

In Program N there exists a number of additional (not standard) tags for
AIML that drive a set of numbers which reflect dimensions in emotions.
Jacco Bikker, who wrote the original version (for a pocket pc) which I
enhanced greatly, invented this emotion engine.  The feature is mostly
parameter driven so it can be tuned to different personalities.  It can be
loaded and unloaded and modified in the scripting language.  When using this
engine the list of responses in a <random> tag can be warped towards
favoring the current emotional state of the bot.

Also available for the cost of waiting until the calculations are performed,
is the interface to ConceptNet.  That tool will take in a text string -
could be the client's input - and return a list of emotions each with a
probability factor.  The basic download for AIMLpad (Program N) contains a
script to access ConceptNet and utilize those indicators.

This might be good if the bot was just to reflect the emotions of the
client, but better yet would be to utilize the built-in expert system to
have a set of rules that decides statistically the best emotional response.
For example, if the bot is excited and the client inputs some harsh words,
maybe a belligerent reply is warranted and influenced by modifying the
emotion engine's internal numbers.

Finally, the AIML output can be annotation to make the MS Agent change to
display the emotion.  Stefan Zakarias added the escape coding to change
speech rate and actions of the MS Agent.  To best work this you might want
to extend an MS Agent to perform more emotion actions.

Tutorials for these features can be found at www.aimlpad.com as well as a
download of the desktop editor.  A much better place to download it is
http://sourceforge.net/projects/program-n/

Happy experimenting,
  Gary Dubuque, playing with music synthesizers to fill in the backgrounds
(like maybe a laugh track)

-----Original Message-----
From: alicebot-general-bounces@...
[mailto:alicebot-general-bounces@...]On Behalf Of Nicola
Alexander Schlup
Sent: Wednesday, January 24, 2007 8:15 AM
To: alicebot-general@...
Subject: [alicebot-general] Natural Language Processing and emotions


Hi all

I would like to know if there are any possibilities to determine
emotions for given sentences with Computational Linguistics/Natural
Language Processing. Is there any known solution for this pattern? It
should not just be an AIML file where every sentence got a static
emotion code. I need such a solution for a research paper where I want
to let a chatbot add emotion code to given sentences. The output could
be used for avaters in 3D tools or any other visualization tool. Thanks
a lot.

Kind regards,
Nicola Alexander Schlup
_______________________________________________
This is the alicebot-general mailing list
Reply to alicebot-general@...
Unsubscribe and change preferences at
http://list.alicebot.org/mailman/listinfo/alicebot-general
Learn netiquette at http://www.dtcc.edu/cs/rfc1855.html
Learn to read at http://www.literacy.org/


_______________________________________________
This is the alicebot-general mailing list
Reply to alicebot-general@...
Unsubscribe and change preferences at http://list.alicebot.org/mailman/listinfo/alicebot-general
Learn netiquette at http://www.dtcc.edu/cs/rfc1855.html
Learn to read at http://www.literacy.org/

[mehri]

The web site:

http://www.aimlpad.com/

has been down for quite a while.


--- Gary Dubuque <gdubuque@...> wrote:

> Hi Nicola,
>
> In Program N there exists a number of additional
> (not standard) tags for
> AIML that drive a set of numbers which reflect
> dimensions in emotions.
> Jacco Bikker, who wrote the original version (for a
> pocket pc) which I
> enhanced greatly, invented this emotion engine.  The
> feature is mostly
> parameter driven so it can be tuned to different
> personalities.  It can be
> loaded and unloaded and modified in the scripting
> language.  When using this
> engine the list of responses in a <random> tag can
> be warped towards
> favoring the current emotional state of the bot.
>
> Also available for the cost of waiting until the
> calculations are performed,
> is the interface to ConceptNet.  That tool will take
> in a text string -
> could be the client's input - and return a list of
> emotions each with a
> probability factor.  The basic download for AIMLpad
> (Program N) contains a
> script to access ConceptNet and utilize those
> indicators.
>
> This might be good if the bot was just to reflect
> the emotions of the
> client, but better yet would be to utilize the
> built-in expert system to
> have a set of rules that decides statistically the
> best emotional response.
> For example, if the bot is excited and the client
> inputs some harsh words,
> maybe a belligerent reply is warranted and
> influenced by modifying the
> emotion engine's internal numbers.
>
> Finally, the AIML output can be annotation to make
> the MS Agent change to
> display the emotion.  Stefan Zakarias added the
> escape coding to change
> speech rate and actions of the MS Agent.  To best
> work this you might want
> to extend an MS Agent to perform more emotion
> actions.
>
> Tutorials for these features can be found at
> www.aimlpad.com as well as a
> download of the desktop editor.  A much better place
> to download it is
> http://sourceforge.net/projects/program-n/
>
> Happy experimenting,
>   Gary Dubuque, playing with music synthesizers to
> fill in the backgrounds
> (like maybe a laugh track)
>
> -----Original Message-----
> From: alicebot-general-bounces@...
>

[mailto:alicebot-general-bounces@...]On

> Behalf Of Nicola
> Alexander Schlup
> Sent: Wednesday, January 24, 2007 8:15 AM
> To: alicebot-general@...
> Subject: [alicebot-general] Natural Language
> Processing and emotions
>
>
> Hi all
>
> I would like to know if there are any possibilities
> to determine
> emotions for given sentences with Computational
> Linguistics/Natural
> Language Processing. Is there any known solution for
> this pattern? It
> should not just be an AIML file where every sentence
> got a static
> emotion code. I need such a solution for a research
> paper where I want
> to let a chatbot add emotion code to given
> sentences. The output could
> be used for avaters in 3D tools or any other
> visualization tool. Thanks
> a lot.
>
> Kind regards,
> Nicola Alexander Schlup
> _______________________________________________
> This is the alicebot-general mailing list
> Reply to alicebot-general@...
> Unsubscribe and change preferences at
>

http://list.alicebot.org/mailman/listinfo/alicebot-general

> Learn netiquette at
> http://www.dtcc.edu/cs/rfc1855.html
> Learn to read at http://www.literacy.org/
>
>
> _______________________________________________
> This is the alicebot-general mailing list
> Reply to alicebot-general@...
> Unsubscribe and change preferences at
>

http://list.alicebot.org/mailman/listinfo/alicebot-general
> Learn netiquette at
> http://www.dtcc.edu/cs/rfc1855.html
> Learn to read at http://www.literacy.org/
>


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com 
_______________________________________________
This is the alicebot-general mailing list
Reply to alicebot-general@...
Unsubscribe and change preferences at http://list.alicebot.org/mailman/listinfo/alicebot-general
Learn netiquette at http://www.dtcc.edu/cs/rfc1855.html
Learn to read at http://www.literacy.org/

[Gary Dubuque]

Yes, a couple of weeks ago our power was out for more than a week.  It got
really cold here.  We were living off our home generator and my web sites
wasn't on that circuit.  I think last week it was out of action for a better
part of a day too.  It has a limit on the number of concurrent accesses that
is pretty low.  Someday I'll switch to the LAMP stack and then be just at
the mercy of my ISP.  I am kind of out in the country here so the
www.aimlpad.com site is pretty sloooow.  No cable in these woods!  Even a
little rush could overwhelm it.  Oh and the virus scanning software might
fustrate you too.  Good thing SourceForge is not that bad!

-----Original Message-----
From: alicebot-general-bounces@...
[mailto:alicebot-general-bounces@...]On Behalf Of mehri
Sent: Thursday, January 25, 2007 9:40 PM
To: Alicebot and AIML General Discussion
Subject: Re: [alicebot-general] [Alicebot-general] Natural
LanguageProcessing and emotions



The web site:

http://www.aimlpad.com/

has been down for quite a while.


--- Gary Dubuque <gdubuque@...> wrote:

> Hi Nicola,
>
> In Program N there exists a number of additional
> (not standard) tags for
> AIML that drive a set of numbers which reflect
> dimensions in emotions.
> Jacco Bikker, who wrote the original version (for a
> pocket pc) which I
> enhanced greatly, invented this emotion engine.  The
> feature is mostly
> parameter driven so it can be tuned to different
> personalities.  It can be
> loaded and unloaded and modified in the scripting
> language.  When using this
> engine the list of responses in a <random> tag can
> be warped towards
> favoring the current emotional state of the bot.
>
> Also available for the cost of waiting until the
> calculations are performed,
> is the interface to ConceptNet.  That tool will take
> in a text string -
> could be the client's input - and return a list of
> emotions each with a
> probability factor.  The basic download for AIMLpad
> (Program N) contains a
> script to access ConceptNet and utilize those
> indicators.
>
> This might be good if the bot was just to reflect
> the emotions of the
> client, but better yet would be to utilize the
> built-in expert system to
> have a set of rules that decides statistically the
> best emotional response.
> For example, if the bot is excited and the client
> inputs some harsh words,
> maybe a belligerent reply is warranted and
> influenced by modifying the
> emotion engine's internal numbers.
>
> Finally, the AIML output can be annotation to make
> the MS Agent change to
> display the emotion.  Stefan Zakarias added the
> escape coding to change
> speech rate and actions of the MS Agent.  To best
> work this you might want
> to extend an MS Agent to perform more emotion
> actions.
>
> Tutorials for these features can be found at
> www.aimlpad.com as well as a
> download of the desktop editor.  A much better place
> to download it is
> http://sourceforge.net/projects/program-n/
>
> Happy experimenting,
>   Gary Dubuque, playing with music synthesizers to
> fill in the backgrounds
> (like maybe a laugh track)
>
> -----Original Message-----
> From: alicebot-general-bounces@...
>

[mailto:alicebot-general-bounces@...]On

> Behalf Of Nicola
> Alexander Schlup
> Sent: Wednesday, January 24, 2007 8:15 AM
> To: alicebot-general@...
> Subject: [alicebot-general] Natural Language
> Processing and emotions
>
>
> Hi all
>
> I would like to know if there are any possibilities
> to determine
> emotions for given sentences with Computational
> Linguistics/Natural
> Language Processing. Is there any known solution for
> this pattern? It
> should not just be an AIML file where every sentence
> got a static
> emotion code. I need such a solution for a research
> paper where I want
> to let a chatbot add emotion code to given
> sentences. The output could
> be used for avaters in 3D tools or any other
> visualization tool. Thanks
> a lot.
>
> Kind regards,
> Nicola Alexander Schlup
> _______________________________________________
> This is the alicebot-general mailing list
> Reply to alicebot-general@...
> Unsubscribe and change preferences at
>

http://list.alicebot.org/mailman/listinfo/alicebot-general

> Learn netiquette at
> http://www.dtcc.edu/cs/rfc1855.html
> Learn to read at http://www.literacy.org/
>
>
> _______________________________________________
> This is the alicebot-general mailing list
> Reply to alicebot-general@...
> Unsubscribe and change preferences at
>

http://list.alicebot.org/mailman/listinfo/alicebot-general
> Learn netiquette at
> http://www.dtcc.edu/cs/rfc1855.html
> Learn to read at http://www.literacy.org/
>


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
_______________________________________________
This is the alicebot-general mailing list
Reply to alicebot-general@...
Unsubscribe and change preferences at
http://list.alicebot.org/mailman/listinfo/alicebot-general
Learn netiquette at http://www.dtcc.edu/cs/rfc1855.html
Learn to read at http://www.literacy.org/


_______________________________________________
This is the alicebot-general mailing list
Reply to alicebot-general@...
Unsubscribe and change preferences at http://list.alicebot.org/mailman/listinfo/alicebot-general
Learn netiquette at http://www.dtcc.edu/cs/rfc1855.html
Learn to read at http://www.literacy.org/