RE: AIML? from John Flynn on 2018-06-27 (semantic-web@w3.org from June 2018)

From: John Flynn <jflynn12@verizon.net>
Date: Wed, 27 Jun 2018 10:05:14 -0400
To: <paoladimaio10@googlemail.com>, "'ProjectParadigm-ICT-Program'" <metadataportals@yahoo.com>
Cc: "'brandon whitehead'" <brandonnodnarb@gmail.com>, "'Eric Prud'hommeaux'" <eric@w3.org>, "'semantic-web at W3C'" <semantic-web@w3c.org>, "'John Flynn'" <jflynn12@verizon.net>
Message-ID: <00f801d40e1f$df648970$9e2d9c50$@net>
I have been interested in the integration of AIML and Semantic Web technology for over a decade. I have submitted several proposals (teaming with AIML developers) to US government agencies to further this research, but discouragingly have received no awards. Semantic Web technology, and specifically OWL, offers great promise in allowing a revolutionary expansion of chat bot capabilities. Rather that storing basically canned responses to queries, an OWL based knowledge base would allow retrieval of class and property data (entities) relevant to the user's query. OWL knowledge bases are normally constructed based on a domain of interest. The more extensive the amount of detail in the knowledge base dictates an increasingly narrower domain of interest just because of the volume of information. This is different from many chat bots that attempt to cover any area that a user might wish to discuss. An AIML-OWL chat bot would need (at least initially to limit the domain of discussion) There are various approaches to OWL knowledge base development but I strongly favor the use of an upper ontology that address all the high-level ontology issues and information that will be used across multiple domains of interest, such a time, space, geography, etc. So, the basic idea is that extensions to AIML natural language parsing capabilities would extract key class or property domain terms from the user input and the system would then search the OWL knowledge base for these concepts. If there were a "hit" then the smart chat bot would then provide the opportunity to increase its capabilities to greatly expand the discussion base because the knowledge base may have a wealth of information related to the concepts in the user's dialog. Additionally, one of the really strong features of an OWL knowledge base is that it provides the capability to conduct reasoning whereby information may be generated via reasoning that is not explicitly present in the knowledge base.  
 
To be honest one of the main reasons I focused on AIML as the front end of such a system is that when this all started I was particularly interested in chat bots (and I still am) and AIML provided a reasonable natural language front end. There have been significant advancements in the area of natural language front ends since those days and any prudent proposal would now need to examine the viability of using something other than AIML. 
 
John Flynn
http://semanticsimulations.com 
 
From: paoladimaio10@gmail.com [mailto:paoladimaio10@gmail.com] On Behalf Of Paola Di Maio
Sent: Wednesday, June 27, 2018 4:22 AM
To: ProjectParadigm-ICT-Program
Cc: paoladimaio10@googlemail.com; brandon whitehead; Eric Prud'hommeaux; semantic-web at W3C
Subject: Re: AIML?
 
Thank you for sharing Milton
 
nice pointer to this interesting project  but, isnt markup languages still
required/useful for representation even in knowledge graphs?
 
and if not, isnt markup language the most basic way to enable intelligent knowledge exchange on the web
so that it can be useful even without a knowledge graph?
 
pdm
 
 
 
On Tue, Jun 26, 2018 at 9:22 AM, ProjectParadigm-ICT-Program <metadataportals@yahoo.com> wrote:
Modeling AI on the web in my humble opinion is no longer a question of simple markup languages, the Internet of Things or more succinctly Internet of Data, Devices, DNA and Digital Agents (IOD4) increasingly uses both AI and virtual reality technologies.
 
In such a setting using ontologies, parsers and any automated process that codes or decodes and interfaces, either in NL setting or otherwise must use of category theory to create the required abstraction for knowledge graphs.
 
Take a look at the Blue Brain Nexus for lateral thinking:
 
BlueBrain/nexus <https://github.com/BlueBrain/nexus> 
 



  


  <https://s.yimg.com/nq/storm/assets/enhancrV2/23/logos/github.png> 

BlueBrain/nexus

nexus - Blue Brain Nexus - A knowledge graph for data-driven science
 
 
Milton Ponson
GSM: +297 747 8280
PO Box 1154, Oranjestad
Aruba, Dutch Caribbean
Project Paradigm: Bringing the ICT tools for sustainable development to all stakeholders worldwide through collaborative research on applied mathematics, advanced modeling, software and standards development
 
On Saturday, May 12, 2018 10:29 AM, Paola Di Maio <paola.dimaio@gmail.com> wrote:
 
Thanks a lot 
 
looks useful- and a good start
 
so MLschema is in practice a... MLML.?
a draft -
 
is it suggested that all knowledge schemas to support
automated reasoning and AI should/could adopt its core elements as its base
(in which I could think if/how this can help my task maybeevaluate it against our use cases)
 
I checked out OpenML and found no credits, who did it, when etc
also it is not clear if its openmarkuplanguage or openmedialibrary
since both seem associated with the same group (Kronos? are they associated with OKF ?)
 both come up in searches
 
 
 
On Sat, May 12, 2018 at 6:58 PM, brandon whitehead <brandonnodnarb@gmail.com> wrote:
Paola,
The machine learning community group [1] published a draft core schema
about a year ago that, at the very least, may be of  interest (link on
main page).

[1] https://www.w3.org/community/ ml-schema/ <https://www.w3.org/community/ml-schema/> 

Cheers,
/Brandon

On 12/05/18 11:51, Paola Di Maio wrote:
> Eric
> Yes, of course getting the key stakeholders involved-
> 
>  since you are familiar with the member base
> i ll be happy to pitch directly members who are working on AI
> if you could suggest a way to shortlist them/approach them
> 
> P
> 
> On Sat, May 12, 2018 at 3:36 PM, Eric Prud'hommeaux <eric@w3.org
> <mailto:eric@w3.org>> wrote:
> 
>     * Paola Di Maio <paola.dimaio@gmail.com
>     <mailto:paola.dimaio@gmail.com >> [2018-05-12 15:15+0530]
>     > Thank you Eric
>     > 
>     > at this stage, I was thinking of some web based knowledge representation
>     > mechanism or ML for something that I am working on related to AI
>     > (I have learned my lessons- glad to share details of this early concept
>     > offlist to those who may express interest until it's solid )
>     > 
>     > I did a search, and found AIML which seems the closest to what I require
>     > however could not find a formal specification to study it further, and
>     > wondered about any interest to W3C.
>     > 
>     > I am pretty sure the web needs what I am thinking of, to what extent its
>     > feasible or we can find folks to do it and adopt it, I dunno
> 
>     To motivate standardization, you have to dig up use cases that not
>     only need some technology, but motivate distinct entities using a
>     common form or interface to that technology. So a win would be
>     e.g. when folks can combine commodity tools to generate such data
>     with commodity tools which make use of it.
> 
> 
>     > >  There's nothing saying you
>     > > can't have a hybrid system which e.g. uses SemWeb for entity
>     > > recognition (à la NCBO annotator) or records ML assertions in
>     RDF for
>     > > further rule execution. That requires people to have expertise and
>     > > commitment in both camps and so far, those folks haven't banded
>     > > together with a set of shared use cases and goals.
>     >
>     >
>     > Am thinking of something fluid,  ML should be sufficient for my
>     requirement
>     > at this stage- also confess that i favour simplicity over
>     sophistication
>     >
>     > but could not find anything that does what I require so thinking maybe
>     > something can be done-
>     >
>     >
>     > > If you can muster
>     > > the troops (an army of five, to be exact), you can easily create
>     a W3C
>     > > Community Group (see [CREATE A COMMUNITY GROUP] at
>     > > <https://www.w3.org/community/ groups/ <https://www.w3.org/community/groups/> 
>     <https://www.w3.org/community/ groups/ <https://www.w3.org/community/groups/> >>).

>     > >
>     >
>     > yep, done it before. I chaired a nice group that did good work for
>     one year
>     > then suddenly fell silent and I am still traumatized  by the
>     experience. :-)
>     > (joke - it was valuable!)
>     >
>     >  anyone interested in AI ML of sorts who is reliable and
>     competent  (not
>     > afraid of failure?) welcome to brainstorm offlist to discuss early
>     stage
>     > concept for this work
>     >
>     > , I need specifically folks who can do implementation side of things
>     > (writing a parsers for validation, and implement the test cases
>     etc) and
>     > who are good at getting research funding - I am okay with the
>     concept and
>     > system design part, and that's about it
>     >
>     > >
>     > > >
>     > >
>     > > The tutorial seemed to be about a template language for natural
>     > > language interfaces while the overview seemed to go more into the
>     > > actual processing logic. Do you know if AIML captures AI logic and
>     > > what use cases would motivate favoring such a standard for Semantic
>     > > Web work?
>     > >
>     >
>     > No, guess not but not sure. AIML seems very very thin at the moment,
>     > although there is a free working prototype online which seems to
>     be using it
>     > https://home.pandorabots.com/ en/ <https://home.pandorabots.com/en/>  <https://home.pandorabots.com/ en/ <https://home.pandorabots.com/en/> >
>     >
>     > I think there's work to be done-
>     >
>     > >
>     > >
>     > > --
>     > > -ericP
>     > >
>     > > office: +1.617.599.3509
>     > > mobile: +33.6.80.80.35.59
>     > >
>     > > (eric@w3.org <mailto:eric@w3.org>)
>     > > Feel free to forward this message to any list for any purpose
>     other than
>     > > email address distribution.
>     > >
>     > > There are subtle nuances encoded in font variation and clever layout
>     > > which can only be seen by printing this message on high-clay paper.
>     > >
>     >
>     >
>     >
>     > --
>     > *A bit about me <https://about.me/paoladimaio> *
> 
>     -- 
>     -ericP
> 
>     office: +1.617.599.3509
>     mobile: +33.6.80.80.35.59
> 
>     (eric@w3.org <mailto:eric@w3.org>)
>     Feel free to forward this message to any list for any purpose other than
>     email address distribution.
> 
>     There are subtle nuances encoded in font variation and clever layout
>     which can only be seen by printing this message on high-clay paper.
> 
> 
> 
> 
> -- 
> *A bit about me <https://about.me/paoladimaio> *
> 



 
-- 
A bit about me <https://about.me/paoladimaio> 
  <https://docs.google.com/uc?export=download&id=0B5XeDqbLLqHSdFVDb0FGMm5vMVE&revid=0B5XeDqbLLqHSOVVJc09yeXI2SlY5OGNhTzdtVTJUUFc5UWh3PQ> 
 



 
-- 
A bit about me <https://about.me/paoladimaio> 
  <https://docs.google.com/uc?export=download&id=0B5XeDqbLLqHSdFVDb0FGMm5vMVE&revid=0B5XeDqbLLqHSOVVJc09yeXI2SlY5OGNhTzdtVTJUUFc5UWh3PQ>
Received on Wednesday, 27 June 2018 14:05:47 UTC