Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Semantic Knowledge and Privacy in the Physical Web

271 views

Published on

In the past few years, the Internet of Things has started to become a reality; however, its growth has been hampered by privacy and security concerns. One promising approach is to use Semantic Web technologies to mitigate privacy concerns in an informed, flexible way. We present CARLTON, a framework for managing data privacy for entities in a Physical Web deployment using Semantic Web technologies. CARLTON uses context-sensitive privacy policies to protect privacy of organizational and personnel data. We provide use case scenarios where natural language queries for data are handled by the system, and show how privacy policies may be used to manage data privacy in such scenarios, based on an ontology of concepts that can be used as rule antecedents in customizable privacy policies.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

Semantic Knowledge and Privacy in the Physical Web

  1. 1. Semantic Knowledge and Privacy in the Physical Web PRAJIT KUMAR DAS, ABHAY KASHYAP, GURPREET SINGH, CYNTHIA MATUSZEK, TIM FININ, ANUPAM JOSHI UMBC ebiquity and IRAL Labs
  2. 2. Motivation Our goal is to provide contextually aware information, using the IoT, that is privacy preserving and ubiquitously helpful Image courtesy Batman Wikia CARLTON Slide 2 of 44
  3. 3. IoT by Volume Slide 3 of 44
  4. 4. IoT by Domain Slide 4 of 44
  5. 5. IoT by Privacy Concerns Slide 5 of 44
  6. 6. Salient features  CARLTON: A context-aware, NL question-answer BOT  Context derived from the Physical Web (IoT)  Under development, prototype system  Simple NLP using tools like Stanford CoreNLP  Mobile app and Kiosk for front-end  ABAC privacy model, Privacy rules using SWRL  Hierarchical context ontology  Optional authentication for UMBC people Slide 6 of 44
  7. 7.  Concretization of IoT  Small, quick seamless interactions with physical objects and locations with your device Physical web: What? Slide 7 of 44
  8. 8. Physical web: What?  Everything is a tap away Slide 8 of 44
  9. 9. Physical web: What?  See what’s useful around you Slide 9 of 44
  10. 10. Physical web: What?  Any object or place can broadcast content Slide 10 of 44
  11. 11. Physical web: How? Three main techniques  Nearby Connections Slide 11 of 44
  12. 12. Physical web: How? Three main techniques  Nearby Connections  Nearby Notifications Slide 12 of 44
  13. 13. Physical web: How? Three main techniques  Nearby Connections  Nearby Notifications  Nearby Messages Slide 13 of 44
  14. 14. System Overview Slide 14 of 44
  15. 15. System Overview Slide 15 of 44
  16. 16. System Overview Slide 16 of 44
  17. 17. Who is Tim Finin? System Overview Slide 17 of 44
  18. 18. “Tim Finin”: Person Entity type “Who”: WH query type Text to Semi-Structured Text Intent Who is Tim Finin? System Overview Slide 18 of 44
  19. 19. “Tim Finin”: Person Entity type “Who”: WH query type Text to Semi-Structured Text Intent SPARQL query generator Context Who is Tim Finin? System Overview Slide 19 of 44
  20. 20. “Tim Finin”: Person Entity type “Who”: WH query type Text to Semi-Structured Text Intent SPARQL query generator Context Who is Tim Finin? Inference Engine OntologyKnowledge base System Overview Slide 20 of 44
  21. 21. “Tim Finin”: Person Entity type “Who”: WH query type Text to Semi-Structured Text Intent SPARQL query generator Context Who is Tim Finin? Inference EngineResponse: JSON {“text”: “He’s a Professor in the Computer Science department!”} OntologyKnowledge base System Overview Slide 21 of 44
  22. 22. Example query Slide 22 of 44
  23. 23. Is this room booked from 2PM-3PM? Example query Slide 23 of 44
  24. 24. User is a faculty and is in front of Conf. room 1. Is this room booked from 2PM-3PM? Example query Slide 24 of 44
  25. 25. Conf. room 1 calendar has no events during that time. Is this room booked from 2PM-3PM? Example query Slide 25 of 44
  26. 26. Is this room booked from 2PM-3PM? No, would you like me to book it from 2PM – 3PM? Example query Slide 26 of 44
  27. 27. Is this room booked from 2PM-3PM? No, would you like me to book it from 2PM – 3PM? Yes, please! Example query Slide 27 of 44
  28. 28. Okay, the room has been booked in your name from 2PM – 3PM Is this room booked from 2PM-3PM? No, would you like me to book it from 2PM – 3PM? Yes, please! Example query Slide 28 of 44
  29. 29. Example query Slide 29 of 44
  30. 30. Is Dr. Joshi here? Example query Slide 30 of 44
  31. 31. Is Dr. Joshi here? User could mean Dr. A. Joshi or Dr. K. Joshi. Example query Slide 31 of 44
  32. 32. Is Dr. Joshi here? But user is in front of Dr. A. Joshi’s office. Example query Slide 32 of 44
  33. 33. Is Dr. Joshi here? User is an advisee of Dr. A. Joshi Example query Slide 33 of 44
  34. 34. Is Dr. Joshi here? Dr. Joshi is in a meeting till 3PM Example query Slide 34 of 44
  35. 35. Example query Slide 35 of 44
  36. 36. Example query Where is Dr. Finin’s office? Slide 36 of 44
  37. 37. Example query User is in CSEE building Where is Dr. Finin’s office? Slide 37 of 44
  38. 38. Example query User is unknown to system Where is Dr. Finin’s office? Slide 38 of 44
  39. 39. Example query Where is Dr. Finin’s office? Please see CSEE front desk for required information Slide 39 of 44
  40. 40. Example1. @prefix crltn:<https://www.ebiquity.org/ontologies/carlton/0.1>. @prefix swrlb:<http://www.w3.org/2003/11/swrlb>. crltn: student(?requester)∧ ( crltn: supervises(“Xavier”,?requester)∨ (crltn: affiliatedWith(?requester,?labName)∧crltn: leads(“Xavier”,?labName)) )∧ crltn: hasCurrentLocation(?requester,?aBldgLocation)∧ crltn: room(?aBldgLocation)∧crltn: sitsIn(“Xavier”,?aBldgLocation)∧ crltn: currentTime(?currTime)∧swrlb: Exists(?anEvent)∧crltn: speakingAt(“Xavier”,?anEvent)∧ ( (crltn: startTime(?anEvent,?eventStartTime)∧swrlb: greaterThan(?eventStartTime,?currTime))∨ (crltn: endTime(?anEvent,?eventEndTime)∧swrlb: greaterThan(?currTime,?eventEndTime)) )∧crltn: hasCurrentLocation(“Xavier”,?aLocation)∧crltn: Location(?aLocation)∧ crltn: requestLocation(“Xavier”) ⇒ shareLocation(?aLocation) Policy Example Slide 40 of 44
  41. 41. supervises(“Xavier”,?requester) OR ( affiliatedWith(?requester,?labName) AND leads(“Xavier”,?labName) ) Policy Example Slide 41 of 44
  42. 42. hasCurrentLocation(?requester,?aBldgLocation) AND room(?aBldgLocation) AND sitsIn(“Xavier”,?aBldgLocation) => shareLocation(?aLocation) Policy Example Slide 42 of 44
  43. 43. Future work  Prototype system constantly adding conversations  Beacons on robots  Reason over robots near you  How robots respond to instructions?  “Can you take me to Prof. Matuszek now?”  “Show me the way to the ITE 346 conference room” Slide 43 of 44
  44. 44. Summary  We presented CARLTON  A context-aware, NL question-answer BOT  Context derived from the Physical Web (IoT)  Semantic web technologies used to preserve data privacy Thanks to NSF for the travel grant! and Thanks to Google for the gift of beacons! Slide 44 of 44

×