NAL-7 as an implicit time intension.

[djire...@gmail.com]

Hi everyone, 

We've been using NARS (our own implementation - matching closely with OpenNARS) and feeding it a series of events. Rather than using NAL-7 we've found it's better to encode events with an explicit time-epoch intension. From NAL-1 the system starts to derive the possible causations and incidentals. For example:

<{ACTIVITY1} --> (& MORNING, 10-01-2018)>

<{ACTIVITY2} --> (& MORNING, 10-01-2018)>

...

<(& MORNING, 10-01-2018) --> [ATTR1]>

|- 

...

<{ACTIVITY1} --> [ATTR1]>

<{ACTIVITY1} --> $X> ==> <{ACTVITIY2} --> $X>>

<{ACTIVITY2} --> $X> ==> <{ACTVITIY1} --> $X>>

(&& <{ACTIVITY1} --> $X>, <{ACTIVITY2} --> $X>

Of course this means we have to include a time-period intension to each relevant statement and question. I suspect this might limit what more knowledge could be inferred from knowledge that is eternal. 

This approach did get me wondering whether NAL-7 could be implemented this way but with an implicit intension.

An event has an implicit intension (which denotes the time epoch) and inferences are done as above. Seemingly, such events could be used with other knowledge that might not have a time component - explicit or implicit. 

To me, it seems a natural extension to the NAL layers but, have I missed something?

Is there a "fly in the ointment?"

I would appreciate any comments. 

Thanks. 

Regards,

David

[Pei Wang]

Hi David,

It is fine to use time-relevant concepts (including relations) like "the morning of 10-01-2018", "after 5 minutes", "when the weather is warm", and so on. This type of temporal reasoning is already possible below NAL-7, except that no time-specific inference will be directly supported by the inference rules, so all the relevant information needs to be provided explicitly in these concepts.

Regards,

Pei

Virus-free.www.avast.com

--
You received this message because you are subscribed to the Google Groups "open-nars" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-nars+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-nars/8b2fc390-5017-45aa-8116-82b470dc1a47n%40googlegroups.com.