Re: deontic logic, deadline coming up

Owen

thanks a lot verbage welcome - will texture it into the paper
//anyone else adding anything to this paper, please initial your paragraph
so that contributorions can be tracked, thanks!//
Owen- what I have in mind (even cursorily) is to show how stratml can
support specific
constructs. Note that I have experience working with expert systems with
rules, where the logic is quite simple, include, exclude, exception and
sometimes some combinations - I have limited experience with complex
rulesets, so this is new also for me


 For example, look at this paper  if you manage to read through the puffed
up language, the bottom line in essence shows that
https://dl.acm.org/doi/10.1145/2514601.2514608

a set of rules (obligations (maintenance, punctual, preemptive) as well as
exceptions (deffeasible obligations) can be implemented using a specific
technique here called
ASP - some examples are given and some verification offered through
a control set called deontic inconsistency.

I do not suggest that we use the same puffed up language, I am personally
inclined to use simple language, although I can boff up my writing if I set
my mind to it

But it should not be dificult to carry out a similar exercise with stratml?

lets start by following the same exact case (could be done with different
cases as well)
using a set of rules that we cal obligations - must do
is there anything in stratml that can be used to represent obligations
(goals
is a bit loose, it shows a sense of direction for what needs to be achieved
but does not show if some action is mandatory)

if not, maybe we need to either find another example (not obligations) or
would need to extend stratML accordingly - the paper could explore either
path

if yes, however, stratml can be used to represent obligations   we need to
figure if these obligations can be represented by stratml to include
subsets (say for example maintanenca, punctual, preemptive etc) as well as
exceptions (cases where such obligations are not mandatory)

we would then finally need to provide some verification mechanism, and
could try to adopt the same (show that stratml can represent the
compliance/measurement of obligations correctly by
Maintenance obligations. Maintenance obligations require a con-dition to
obtain during all instants before a deadline. For instance,(from [21])
after opening a bank account, customers must keep apositive balance until
bank charges are taken out. The effect ofactionopening_accountis modeled by
the action law:[opening_account]O(pos_balanceUcharges_taken_out)A
maintenance obligation persists until the deadline is reached orthe
obligation is violated or cancelled. The dynamic of mainte-nance
obligations is ruled by persistency, fulfillment and violationlaws for
until obligations in Section 5.Punctual obligations. They can be modeled by
obligations ofthe formO(Xl). Suppose, e.g., that when a system receives agiven
message, it must immediately acknowledge it:O(Xack)←received_message.
According to the rules in Section 5, if notcanceled, this obligation causes
the obligationO(ack)to be addedto the next state.Preemptive and
nonpreemptive obligations. [21] distinguishespreemptive and non-preemptive
achievement obligations. Considerthe following example: after a contract
has been signed, a copy hasto be sent within a given deadline. If the copy
has been sent beforethe contract has been signed, this does not fulfill the
obligation. Inthis case, the obligation is said to benon-preemptive. To
capturenon-preemptive obligations, we introduce an obligation to executean
action as follows:O(¬deadlineU〈send_copy〉>). This obli-gation is fulfilled
if the actionsend_copyis executed within thedeadline (and after the
obligation has been generat

*The verification that a business process is compliant with a set
ofbusiness rules *
*consists in verifying that all the norms or business rules are satisfied
in all the execution*
* of the process.*

If we can do that, we have a paper - we only need to write it
if we cannot do that, we may have to focus the paper more abstract way to
natural language examples



On Sun, Mar 8, 2020 at 12:07 AM Owen Ambur <Owen.Ambur@verizon.net> wrote:

> Paola, in light of the deadlines, I'm not sure we'll have time to produce
> an output of high enough quality.  I also wonder whether the program
> committee is open to considering the points I believe are pertinent.
>
> However, I took a shot at including some verbiage in your draft at
> https://docs.google.com/document/d/1UovxH61OEHr33QdVR53Kcs0mJPQc3V4cE95i-QyqrQ8/edit
>
> Owen
> On 3/7/2020 2:17 AM, Paola Di Maio wrote:
>
> deadline extended
>    Abstract Submission Deadline:  March 8th March 15th, 2020
> •    Paper Submission Deadline:  March 15th March 22nd, 2020
>
> On Sat, Mar 7, 2020 at 7:38 AM Paola Di Maio <paola.dimaio@gmail.com>
> wrote:
>
>> Owen
>> since you are familiar with stratML better than any of us, perhaps
>> after reading the intro on Standford website (to help you figure the jist
>> of deonticL at a glance) you could perhaps write two or three paragraphs to
>> explain
>> how stratML can be used to implement deontic logic, as well as its
>> limitations
>> (how stratML does not support the constructs) - ideally with a few
>> examples of each
>> instance
>>
>> We could then use it as an an example of how a structured natural
>> language schema
>> can be used to support deontL and add it to the paper
>>
>> The bottom line being to show how natural language KR can be used to
>> enforce/monitor the compliance of norms
>>
>> Let us know what you think
>> PDM
>>
>>
>>
>>
>> On Sat, Mar 7, 2020 at 3:05 AM Owen Ambur <Owen.Ambur@verizon.net> wrote:
>>
>>> Paola, your words are music to my ears.  While I'm not sure I'll have
>>> enough to contribute to be considered a coauthor, I will be happy to add
>>> any value of which I am capable.  For example, it seems to me that
>>> Consequentalism is also highly relevant.
>>> https://plato.stanford.edu/entries/consequentialism/  Of what value are
>>> rules if the consequences of following them are uncertain, unknown, or
>>> worse?
>>>
>>> It seems to me that we have far too many laws, regulations, policies,
>>> directives, and guidance documents in narrative format, which hardly anyone
>>> has the time to read much less comprehend or act upon ... and far too
>>> performance plans -- with clearly specified stakeholder roles and
>>> performance metrics -- in open, standard, machine-readable format, like
>>> StratML.
>>>
>>> BTW, I am reading Jason Brennan's book, *Against Democracy*.  I've just
>>> started but in the preface he asserts: "... politics ... pulls us apart,
>>> stultifies and corrupts us, and makes us civil enemies."  While I may not
>>> agree with his proposed solution, his description of the problem resonates
>>> strongly with me.  It supports the personal biases I voiced in my posting
>>> at
>>> https://www.linkedin.com/pulse/transforming-governance-reducing-cost-gofpau-owen-ambur/
>>>
>>> Owen
>>> On 3/6/2020 4:04 AM, Paola Di Maio wrote:
>>>
>>> It looks to me that AI/KR and StratML are a natural implementation
>>> vehicle for deontic logic, in many ways
>>>
>>> Some literature exists  exploring the association of DeonL and computer
>>> science, the topic is
>>> complex I can understand why the cognitive load required may exceed the
>>> capacity of this group at the moment but never know
>>>
>>> I am now  exploring further the convergence between AI/KL ande Deonl,
>>>
>>> Let me start by introducing some references
>>> https://plato.stanford.edu/entries/logic-deontic/challenges.html
>>> https://plato.stanford.edu/entries/logic-deontic/
>>>
>>>  Based on the intro above, it looks to me that  stratML as understand it
>>> so far, has a potential to implement deontic logic neatly. at least in
>>> part, and if we decide to follow up this avenue, stratML may have to evolve
>>> accordingly (to support deontL constructs not yet supported for example)
>>> but that would enhance it greatly, in the long term
>>>
>>>
>>> I am attempting to draft an abstract with a view to have a paper
>>> completed in one week
>>>
>>> https://docs.google.com/document/d/1UovxH61OEHr33QdVR53Kcs0mJPQc3V4cE95i-QyqrQ8/edit
>>>
>>>
>>> anyone interested to coauthor let me know, if not just wish me good luck
>>>
>>> cheers
>>>
>>> PDM
>>>
>>>
>>>
>>>
>>>

Received on Monday, 9 March 2020 03:06:45 UTC