Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* ·...
Transcript of Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* ·...
![Page 1: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/1.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Antonis Bikakis Department of Informa-on Studies University College London
Dem@Care Summer School on Ambient Assisted Living Chania, Crete, September 2013
![Page 2: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/2.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Outline n Context and Contextual Reasoning in Ambient Intelligence n A Centralized Reasoning Framework n R-‐CoRe – A Distributed Approach n Centralized vs. Distributed Reasoning n Open Problems
![Page 3: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/3.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Ambient Intelligence n Goal: Transform our living and working environments into smart spaces n Requirement: Augment environments with sensing, compu-ng,
communica-on and reasoning capabili-es
![Page 4: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/4.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Context is any informa/on that can be used to characterize the situa/on of an en/ty. An en/ty is a person, place or object that is considered relevant to the interac/on between a user and applica/on, including the user and applica/ons themselves
[Dey and Abowd, 1999]
Context
![Page 5: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/5.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Context RepresentaCon n Key-‐value models
q Service: list of aSributes in a key-‐value manner n Markup scheme models
q XML-‐based n Graphical models
q UML like n Object oriented models
q Context data encapsulated in data objects n Logic-‐based models
q First Order Logic, Logic Programming n Ontology-‐based models
q Based on Descrip-on Logics
![Page 6: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/6.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Contextual Reasoning n Aims
q Inference of high-‐level context knowledge q Consistency checking q Context-‐aware decision making
n Challenges q Imperfect context informa-on q Heterogeneous en--es q Highly dynamic and open environments q Distributed context informa-on q Unreliable wireless communica-ons q …restricted by the range of transmiSers
![Page 7: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/7.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Contextual Reasoning (cont’d) n Approaches
q Ontological reasoning n DL rules used to derive implicit knowledge + Natural integra-on with ontology model – Limited reasoning capabili-es
q Rule-‐based reasoning n More expressive rule languages n FOL, Logic Programming, Defeasible Logic
q Probabilis-c reasoning n Explicit model uncertainty, confidence values, causal rela-onships + Rich expressive capabili-es − High complexity
![Page 8: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/8.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Rule-‐based Contextual Reasoning n Benefits
q Simplicity & Flexibility q Formality q Expressiveness q Modularity q High-‐level abstrac-on & Informa-on hiding q Integra-on with ontology languages
![Page 9: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/9.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Outline n Context and Contextual Reasoning in Ambient Intelligence n A Centralized Reasoning Framework n R-‐CoRe – A Distributed Approach n Centralized vs. Distributed Reasoning n Open Problems
![Page 10: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/10.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Aims & Architecture n Part of a large-‐scale Ambient Intelligence facility developed for the
needs of the ICS-‐FORTH Ambient Intelligence Programme
n Design Goals q efficient representa-on, monitoring, dissemina-on of context q reasoning about the available informa-on q context-‐aware decisions
![Page 11: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/11.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Rule Types n Inference rules
q Triggered by new asser-ons in the KB q Assert new rela-ons in the KB
n Ac-on rules q Reac-ve (to events) or Triggered (by asser-ons in the KB) q Assert new rela-ons in the KB q Determine and send commands for ac-ons
n Rule Scheme (ECA) event(E),/* received from middleware and added as fact in KB */ precondition(C1),…, precondition(Ck) /* relations in KB */ -> action(A1),…, action(An) /* functions for KB update or commands for actions sent to middleware */
![Page 12: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/12.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Special Features n Seamless Interac-on
q Adjust services to user’s context q Achieved through
n Sensing – keep track of user’s context n high-‐level context inference – iden-fy state / situa-on n context-‐aware reasoning – situa-on-‐based policies
n Vast amount of context informa-on q Context Classifica-on q Context Segmenta-on
n Inconsistency Resolu-on q Conflicts due to compe-ng policies q Priority-‐based rule classifica-on
![Page 13: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/13.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Outline n Context and Contextual Reasoning in Ambient Intelligence n A Centralized Reasoning Framework n R-‐CoRe: A Distributed Approach n Centralized vs. Distributed Reasoning n Demo
![Page 14: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/14.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Features & Underlying Technologies n R-‐CoRe: A Rule-‐based Contextual Reasoning Plagorm for AmI n Developed with SnT Luxembourg for the needs of the CoPAInS
(Conviviality and Privacy in Ambient Intelligence Systems) project q Funded by FNR Luxembourg
n Main Features q Distributed q Rule-‐based q Non-‐monotonic q Preference-‐based conflict resoluCon q Dynamic & AdapCve
n Underlying technologies q MulC-‐Context Systems q Contextual Defeasible Logic (CDL) q Kevoree
![Page 15: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/15.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
MulC-‐Context Systems: The magic box example
Mr. 2
Mr. 1 Mr. 1
![Page 16: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/16.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
MulC-‐Context Systems: The magic box example n None of the observers can make
out the depth of the box
Mr. 2
Mr. 1 Mr. 1
![Page 17: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/17.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
MulC-‐Context Systems: The magic box example n None of the observers can make
out the depth of the box n Mr. 1’s beliefs may regard
concepts that are meaningless for Mr.2 and vice versa
Mr. 2
Mr. 1
central section
?
![Page 18: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/18.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
MulC-‐Context Systems: The magic box example n None of the observers can make
out the depth of the box n Mr. 1’s beliefs may regard
concepts that are meaningless for Mr.2 and vice versa
n Mr. 1 and Mr. 2 may use common concepts but interpret them in different ways
Mr. 2
Mr. 1
right section
right section
![Page 19: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/19.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
MulC-‐Context Systems: The magic box example n None of the observers can make
out the depth of the box n Mr. 1’s beliefs may regard
concepts that are meaningless for Mr.2 and vice versa
n Mr. 1 and Mr. 2 may use common concepts but interpret them in different ways
n The observers may have par-al access to each other’s beliefs about the box.
Mr. 2
Mr. 1
ball in the left section
ball in the right
section
![Page 20: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/20.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
MulC-‐Context Systems: IntuiCons and Model
n Context n A parCal and approximate theory of the world from some individual’s
perspecCve n A logical theory – a set of axioms and inference rules
n Mul--‐Context Systems n Distributed context theories connected through mappings that
enable informa-on flow between different contexts n Mappings modeled as inference rules with premises and
consequences in different contexts
![Page 21: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/21.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Nonmonotonic MCS n MCS enriched with nonmonotonic features to handle imperfec-ons, e.g.
incomplete knowledge, inconsistencies
Context C
¬k k Context A
Context B
![Page 22: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/22.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Contextual Defeasible Logic n …
q .. A Defeasible MCS C is a collec-on of contexts Ci
Each context Ci is a tuple (Vi , Ri , Ti ) q Vi : vocabulary used by Ci q Ri : set of rules q Ti : preference ordering on C
Vi : a set of literals of the form a, ¬ a
![Page 23: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/23.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Contextual Defeasible Logic (cont’d)
Three types of rules in Ri
q Strict local rules
ril : (ci : a1),…, (ci : an-1) → (ci : an)
q Defeasible local rules
rid : (ci : a1),…, (ci : an-1) ⇒ (ci : an)
q Mapping rules
rim : (cj : a1),…, (ck : an-1) ⇒ (ci : an)
Ti is a par-al preference ordering on C modeled as a Directed Acyclic Graph
![Page 24: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/24.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
AAL Example Scenario
emergency
prone to heart attack normal pulse
lying on the floor
Ac-vity Recogni-on
Wearable Health Bracelet
Medical Profile
Home Care System
SMS System
![Page 25: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/25.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario in CDL terms
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
![Page 26: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/26.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario in CDL terms
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
rhcsm1 : (br :normalPulse)
⇒ (hcs : ¬ emergency)
rhcsm2 : (arm :lyingOnFloor), (med :proneToHA)
⇒ (hcs : emergency)
![Page 27: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/27.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario in CDL terms
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
rhcsm1 : (br :normalPulse)
⇒ (hcs : ¬ emergency)
rhcsm2 : (arm :lyingOnFloor), (med :proneToHA)
⇒ (hcs : emergency)
rmedl : → (med : proneToHA)
![Page 28: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/28.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario in CDL terms
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
rhcsm1 : (br :normalPulse)
⇒ (hcs : ¬ emergency)
rhcsm2 : (arm :lyingOnFloor), (med :proneToHA)
⇒ (hcs : emergency)
rmedl : → (med : proneToHA)
rbr
l : → (br : normalPulse)
![Page 29: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/29.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario in CDL terms
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
rhcsm1 : (br :normalPulse)
⇒ (hcs : ¬ emergency)
rhcsm2 : (arm :lyingOnFloor), (med :proneToHA)
⇒ (hcs : emergency)
rmedl : → (med : proneToHA)
rbr
l : → (br : normalPulse)
rarml : → (arm : lyingOnFloor)
![Page 30: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/30.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Distributed Query EvaluaCon n When a context receives a query for one of its local literals q
q Evaluates answer based on local knowledge If not possible q Collects relevant informa-on from other contexts through mappings q Checks applicability of rules for and against q q Evaluates answer based on
-‐ Applicable rules -‐ Preferences
![Page 31: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/31.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario: Query EvaluaCon
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
rhcsm1 : (br :normalPulse)
⇒ (hcs : ¬ emergency)
rhcsm2 : (arm :lyingOnFloor), (med :proneToHA)
⇒ (hcs : emergency)
rmedl : → (med : proneToHA)
rbr
l : → (br : normalPulse)
rarml : → (arm : lyingOnFloor)
![Page 32: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/32.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario: Query EvaluaCon
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
rhcsm1 : (br :normalPulse)
⇒ (hcs : ¬ emergency)
rhcsm2 : (arm :lyingOnFloor), (med :proneToHA)
⇒ (hcs : emergency)
rmedl : → (med : proneToHA)
rbr
l : → (br : normalPulse)
rarml : → (arm : lyingOnFloor)
![Page 33: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/33.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario: Query EvaluaCon
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
rhcsm1 : (br :normalPulse)
⇒ (hcs : ¬ emergency)
rhcsm2 : (arm :lyingOnFloor), (med :proneToHA)
⇒ (hcs : emergency)
rmedl : → (med : proneToHA)
rbr
l : → (br : normalPulse)
rarml : → (arm : lyingOnFloor)
![Page 34: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/34.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario: Query EvaluaCon
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
rhcsm1 : (br :normalPulse)
⇒ (hcs : ¬ emergency)
rhcsm2 : (arm :lyingOnFloor), (med :proneToHA)
⇒ (hcs : emergency)
rmedl : → (med : proneToHA)
rbr
l : → (br : normalPulse)
rarml : → (arm : lyingOnFloor)
![Page 35: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/35.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario: Query EvaluaCon
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
rhcsm1 : (br :normalPulse)
⇒ (hcs : ¬ emergency)
rhcsm2 : (arm :lyingOnFloor), (med :proneToHA)
⇒ (hcs : emergency)
rmedl : → (med : proneToHA)
rbr
l : → (br : normalPulse)
rarml : → (arm : lyingOnFloor)
![Page 36: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/36.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario: Query EvaluaCon
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
rhcsm1 : (br :normalPulse)
⇒ (hcs : ¬ emergency)
rhcsm2 : (arm :lyingOnFloor), (med :proneToHA)
⇒ (hcs : emergency)
rmedl : → (med : proneToHA)
rbr
l : → (br : normalPulse)
rarml : → (arm : lyingOnFloor)
![Page 37: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/37.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario: Query EvaluaCon
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
rhcsm1 : (br :normalPulse)
⇒ (hcs : ¬ emergency)
rhcsm2 : (arm :lyingOnFloor), (med :proneToHA) Thcs=[med,arm,br]
⇒ (hcs : emergency)
![Page 38: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/38.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario: Query EvaluaCon
emergency
prone to heart attack normal pulse
lying on the floor
rsmsm : (hcs :emergency)
⇒ (sms : dispatchSMS)
rhcsm1 : (br :normalPulse)
⇒ (hcs : ¬ emergency)
rhcsm2 : (arm :lyingOnFloor), (med :proneToHA) Thcs=[med,arm,br]
⇒ (hcs : emergency)
![Page 39: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/39.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Kevoree n Open source project available at: www.kevoree.org
q Enables distributed reconfigurable sokware development q Any sensor, sokware applica-on, web service can be represented as a
component (with I/O) in Kevoree q The set of services/applica-ons offered by a single en-ty (e.g. device) is
represented as a Kevoree node q Channels represent different types of communica-on among components
(TCP/IP, email, SMS, etc.)
A Kevoree component
Input ports Output ports
![Page 40: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/40.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Kevoree in R-‐CoRe n Each en-ty (mobile compu-ng device) is implemented as a Kevoree
node. n Each context is implemented as a Kevoree component. n Kevoree channels enable exchange of informa-on (messages) between
different components. n Kevoree’s adap-ve and auto-‐discovery capabili-es enable detec-ng
new nodes and adap-ng to any context changes.
![Page 41: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/41.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
R-‐CoRe Architecture
Knowledge Base (Context) Query Servants
Local Rules Mapping rules Preferences
Input: QueryIn ConsoleIn
Output: QueryOut ConsoleOut
Input ports Output ports
![Page 42: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/42.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario: in R-‐CoRe terms Interceptor: Another component we developed to capture and display all the interac-ons (Queries/responses)
Query components: Each one corresponds to the context of a different en-ty
![Page 43: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/43.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Example Scenario: in R-‐CoRe terms (cont’d)
Rule bases and preferences in the example scenario
![Page 44: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/44.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
R-‐CoRe: Demo
n You can download the demo and test it yourself from hSps://github.com/securityandtrust/ruleml13
![Page 45: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/45.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
A Smart Classroom Scenario
classCme classroom
one person detected
no class acCvity
![Page 46: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/46.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
A Social Mobile CompuCng Scenario
User A at university
User B downtown
User C at FORTH
University Bluetooth server
FORTH Bluetooth server
CS585 canceled
CS585 canceled
CS585 canceled
SW lecture tennis tourn.
SW lecture tennis tourn.
SW lecture
![Page 47: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/47.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Outline n Context and Contextual Reasoning in Ambient Intelligence n A Centralized Reasoning Framework n R-‐CoRe: A Distributed Approach n Centralized vs. Distributed Reasoning n Open Problems
![Page 48: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/48.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Centralized vs. Distributed Reasoning n Distribu-on of knowledge n Reasoning with the whole picture n Scalability n Computa-onal Issues
q Single powerful computer vs.
q Devices with limited resources n Communica-on Issues
q Small size of messages vs.
q Small number of messages n Points of failure n Privacy
![Page 49: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/49.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Outline n Context and Contextual Reasoning in Ambient Intelligence n A Centralized Reasoning Framework n R-‐CoRe: A Distributed Approach n Centralized vs. Distributed Reasoning n Open Problems
![Page 50: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/50.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence
Open Problems n Privacy – Security
n Open environments n Unno-ceable access to personal data
n Conviviality n Means and incen-ves for coopera-on n Reconciling conviviality with privacy
n Planning n Common plans n Efficient Plan Execu-on
n Learning n Iden-fy user’s needs and inten-ons n Computa-onal Benefits
n Verifica-on & Valida-on
![Page 51: Rule%based*Contextual*Reasoning*in* Ambient*Intelligence* · Rule%based*Contextual*Reasoning*in*Ambient*Intelligence* MulC%Context*Systems:*The*magicbox example * None!of!the!observers!can!make!](https://reader033.fdocuments.us/reader033/viewer/2022052001/6013845c950682317755b4d0/html5/thumbnails/51.jpg)
Rule-‐based Contextual Reasoning in Ambient Intelligence