Towards a Logical Framework For Knowledge Management in Digital Economy

wildlifeplaincityManagement

Nov 6, 2013 (3 years and 5 months ago)

45 views

1






Heriot

-
Watt University


MACS, Edinburgh



lilia@macs.hw.ac.uk





Towards a Logical Framework For Knowledge
Management in D
igital
E
conomy



Lilia Georgieva and
Imran

Zia


KES
-
AMSTA 2009

June, 5
th




2

Structure



Background and motivation


Knowledge management in digital economy


Formalisation and verification


Case studies



3

Challenges in digital economy



Knowledge generation


Knowledge creation and sharing


Analysis of knowledge management models


Adaptation of knowledge management models


Formalisation of knowledge management


Verification and security


4

Background & Motivation


Languages for knowledge modelling


-
modal logic;


-
description logic;


-
epistemic logic;


-
process logic.


Model checking and verification.



Benefits.


Case studies.




5

Knowledge Management

Knowledge

management

is
:


a

mix

of

contextual

information,

experience

and

rules






an

integrated

approach

to

identifying,

capturing,

evaluating,

retrieving

and

sharing

all

of

enterprises

information

assets
.


Types

of

knowledge

in

digital

economy
:

descriptive,

static,

updated,

dynamic,

interactive
.










6


Example of Static KM Model


Information processing view of knowledge work = limited to static combination
of natural language.

(
Kimas

2003)

7



Our modelling language




Prepositional logic formulas + two (basic) modal logic
operators :




(diamond) possibly




†††††
⡢(砩x湥n敳獡ey



For proposition letters set
{ p, q, r…},

if


,


are modal
formulas, then so are:


T,



Ⱐ,€

,


v

,




,





,





,



,



.





8

Epistemic Logic


Logic for knowledge and belief.


Provides an insight in to properties of individuals and
groups.


Knowledge is implicitly represented in the agent’s
information state.



9

Standard Knowledge Axioms


(K
1
) Ki ¬¬p


Ki p,





(K
2
) ¬Ki¬p

Ki p,





(K
3
)¬Ki p

Ki¬ p,





(K
4
) Ki p


Ki Ki p





(K
5
) Ki p
۸

Ki (p

q)


Ki q,





(K
6
) Ki (p
۸
q)


Ki p
۸

Ki q,





(K
7
) Ki (p
٧
q)


Ki p
٧

Kiq,





(K
8
) Ki ¬ (p
۸
q)


Ki ¬ p
٧

Ki ¬ q,





(K
9
) Ki ¬ (p
٧
q)


Ki ¬ p
۸

Ki ¬ q,





(K
10
) Ki Kj p


Ki p.

10

Knowledge Management Streams


Knowledge sharing


Determine knowledge


Determine knowledge available


Determine knowledge gap


Knowledge lock


Knowledge utilisation


Knowledge evaluation


11

Knowledge management framework

The framework is based on the following steps:



Knowledge
stream specification.


Knowledge stream formalisation in epistemic logic.


Translation into MAP encoding.


Semantic interpretation (FIPA semantics).


Stream verification.


Analysis of stream behaviour.

12

Graphical Representation of Knowledge
Sharing Protocol



Reject

Accept

a
gent j

shares

fact

a
gent
i

knows

fact

Initial

Shares (
M.

SHARES (
M)

SHARES
(M
)

REJECTS
(M
)


REJECTS (
M)

ACCEPT (
M)


ACCEPT
(M
)

Model checking


Input:


A finite state model of the knowledge process.


A specification (requirement) regarding the model.


Algorithm: Check if the system design satisfies the
specification. If not, generate a system run which
violates the specification (counter example).


13

Model checking knowledge
processes


SPIN is a verification tool used for data
communication protocols, multi
-
threaded code,
client
-
server applications.


SPIN verification is focused on proving correctness of
processes interactions; not much importance is given
to internal computations.


Processes refer to components that interact with each
other.


Interaction is synchronous or asynchronous message
passing.


14

15

Knowledge sharing

Initial state



(
Kj

q
) (
Ki

p
)

Generated axioms


Ki
p



Ki Ki
p

(K
4
) Kj
q



Kj Kj
q

(K
4
)

¬
Ki

p


Ki
¬
p

, (K
3
) ¬
Kj

q


Kj
¬
q

, (K
3
)

¬
Ki
¬
p


Ki

p
, (K
2
) ¬
Kj
¬
q


Kj

q

, (K
2
)


Ki

¬ ¬
Kj

q



Ki

q

, (K
1
)

End State


Ki (
p

۸

q
)


Ki
q

۸

Ki
p
, (K
6
)



Example
properties:

Data is always sent from the agent with the latest data to the other agent.


Exactly one agent is the active knowledge sharer at any point of time.












End process

insert end process



16

Related work on KM

Knowledge management models



Static KM models


Computational models


Diagramatic

KM models



Related work on Model Checking



Agent communication


Fairness,
liveness
, termination.


Security of information.

17

Thank you!