On Thu, Oct 13, 2016 at 6:08 PM, Vishnu Priya <vishnupriya...@gmail.com>
> My idea is to give, streams of text ( i don't know what it is about) -- >
> NLP pipeline --> stimulate atoms --> retrieve the atoms with high STI and
> its related atoms and check whether any useful/important info is obtained.
> Regarding this, i would like to clarify and conform few things.
> 1. Can i stimulate atoms every time they have been parsed by the NLP
> pipeline ? Something like this exists before?. I just tried (nlp-parse
> "The sky is beautiful") then (cog-stimulate). But i did not know the exact
> syntax for stimulating. Is this the way it works? What if, when i load a
> scheme file with atoms in AS and want to stimulate? How to ensure that the
> atoms with high STI are remembered/forgotten?.
> Currently, Here are the steps to run the ECAN system with the nlp pipeline
- Go to opencog shell and type loadmodule libattention.so (use full
- Type start-ecan
- Goto guile shell by typing scm
- type (nlp-start-stimulation STIMULUS_AMOUNT) for a moderately
sized atomspace I am using STIMULUS_AMOUNT to have a value of 100. What the
forgetting agent currently does is removing atoms but we never really
worked with it since our main focus was on the spreading, hebbian and rent
collection agent which are responsible in determing the content of the
> 1. Feeding the atoms in AS should be in the form of relex or
> relex2logic for boosting?
> There is no constraint on this. You can feed any way you want as long as
you are stimulating the particular atoms of interest with the
(cog-stimulate atom stimulus) function from scheme or some with an agent
that runs with in the cogserver.
> 1. It is not possible to store atoms with high attention values in
> DB. There is no code for doing this at this time?. I can also never delete
> atoms once that are put into DB.?
> Linas will answer this better.
> 1. I need to write a code, so that i can retrieve top ranking atoms
> with high STI?
> Atoms in the attentional focus could be feteched with the (cog-af)
function in scheme. It returns all atoms with STI values above the
attentional focus boundary STI value. But if you are looking for ranking
atoms based on their STI including atoms outside of the attentional focus,
you will need to write a C++ code which calls this function
with lower and upper boundary STI values provided.
After these steps, everytime you parse a sentence (eg. (nlp-parse "Hello
World!") ) WordNodes and WordInstanceNodes will get stmulated with the
You may view whats in the attentional focus and and and the attentional
focus boundary with (cog-af) and (cog-af-boundary) methods respectively.
There are few AV related scheme functions listed here
> I can give only the sentences which are grammatically correct into NLP
> Pipeline?. So when i give incomplete sentences say human chats like this
> "MadMonkey's Jim Cramer next Weeeeeeeek investment htt brtt", it will be
> parsed with low confidence value. Apart from this, will it have impact on
> providing boost to atoms?
> Thanks and regards,
> You received this message because you are subscribed to the Google Groups
> "opencog" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to opencog+unsubscr...@googlegroups.com.
> To post to this group, send email to email@example.com.
> Visit this group at https://groups.google.com/group/opencog.
> To view this discussion on the web visit https://groups.google.com/d/
> For more options, visit https://groups.google.com/d/optout.
Belachew Misgana Bayetta
'For everyone who asks receives, and the one who seeks finds, and to the
one who knocks it will be opened ' --Jesus
'The future belongs to those who believe in the beauty of their dreams'
You received this message because you are subscribed to the Google Groups
To unsubscribe from this group and stop receiving emails from it, send an email
To post to this group, send email to firstname.lastname@example.org.
Visit this group at https://groups.google.com/group/opencog.
To view this discussion on the web visit
For more options, visit https://groups.google.com/d/optout.