Re: OWA vs CWA

Dear Joshua,

In the problem at hand, we don't enter into this discussion, for the 
Evaluator receives one policy, with one zero or more rules. We don't 
care about the rest of the policies in the universe. The Evaluator has 
to evaluate against the given policy. Nothing else. We have complete 
information (CWA). And we need to decide whether we authorise smoking or 
not with those rules. And there are the two possible behaviours. :)

Regards,
Víctor

El 09/06/2025 a las 15:25, Joshua Cornejo escribió:
>
> Hello,
>
> I was obviously “cold” on my memory cache about the topic during the 
> call, but here is the definition that matches my understanding (from 
> Wikipedia):
>
> “The*closed-world assumption*(*CWA*), in aformal system of logic 
> <https://en.wikipedia.org/wiki/Mathematical_logic>used forknowledge 
> representation 
> <https://en.wikipedia.org/wiki/Knowledge_representation>, is the 
> presumption that a statement that is true is also known to be true. 
> Therefore, conversely, what is not currently known to be true, is 
> false. The same name also refers to alogical 
> <https://en.wikipedia.org/wiki/Formal_logic>formalization of this 
> assumption byRaymond Reiter 
> <https://en.wikipedia.org/wiki/Raymond_Reiter>.^[1] 
> <https://en.wikipedia.org/wiki/Closed-world_assumption#cite_note-1> 
> The opposite of the closed-world assumption is the*open-world 
> assumption*(*OWA*), stating that lack of knowledge does not imply falsity.
>
> Then the summary for OWA/CWA from ChatGPT also matches my points:
>
>   * Use *CWA* when:
>       o You control the data and know it's complete.
>       o Making false positives is worse than false negatives.
>       o You're designing systems like banking rules, logistics, etc.
>   * Use *OWA* when:
>       o You're dealing with incomplete or distributed data.
>       o You need to reason under uncertainty (e.g., AI, law).
>       o You're building ontologies, knowledge graphs, or inference
>         engines.
>
> My thinking when raising the issue is that we are not dealing with any 
> of the 3 cases **_at evaluation_** to justify an OWA (we have a state 
> of the world, which contains all the necessary information and we are 
> not building an KG), while the 3 points: control, false positives & 
> logistics (which is rights management)
>
> Regards,
>
> ___________________________________
>
> *Joshua Cornejo*
>
> *marketdata <https://www.marketdata.md/>*
>
> smart authorisation management for the AI-era
>

-- 
Víctor Rodríguez-Doncel
D2110 - Ontology Engineering Group (OEG)
Departamento de Inteligencia Artificial
ETS de Ingenieros Informáticos
Universidad Politécnica de Madrid
https://cosasbuenas.es

Warning about the confidentiality of communications: the content of this e-mail and its attachments is strictly confidential. Any unauthorized copying, disclosure or distribution of the material in this e-mail is forbidden.