Mobile, Collaborative and Context-Aware Systems

Slides:



Advertisements
Similar presentations
AVATAR: Advanced Telematic Search of Audivisual Contents by Semantic Reasoning Yolanda Blanco Fernández Department of Telematic Engineering University.
Advertisements

CSCI 572 Project Presentation Mohsen Taheriyan Semantic Search on FOAF profiles.
1 Personal Activity Coordinator (PAC) Xia Hong UC Berkeley ISRG retreat 1/11/2000.
Copyright © Norman Sadeh Semantic Web Technologies to Reconcile Privacy and Context Awareness Norman M. Sadeh ISRI- School of Computer Science.
Enterprise Privacy Promises and Enforcement Adam Barth John C. Mitchell.
Audumbar Chormale Advisor: Dr. Anupam Joshi M.S. Thesis Defense
Semantic Web Technologies Lecture # 2 Faculty of Computer Science, IBA.
Architectural Design.
SPARQL All slides are adapted from the W3C Recommendation SPARQL Query Language for RDF Web link:
Anupam Joshi and Tim Finin Ebiquity UMBC
Tim Finin University of Maryland, Baltimore County Joint work with Anupam Joshi, Laura Zavala, Radhika Dharurkar, Pramod Jagtap, Dibyajyoti Ghosh and Amey.
Aegis: A Semantic Implementation of Privacy as Contextual Integrity in Social Ecosystems Imrul Kayes, Adriana Iamnitchi.
Intelligent Agents Meet the Semantic Web in Smart Spaces Harry Chen,Tim Finin, Anupam Joshi, and Lalana Kagal University of Maryland, Baltimore County.
Knowledge based Learning Experience Management on the Semantic Web Feng (Barry) TAO, Hugh Davis Learning Society Lab University of Southampton.
1 A pattern language for security models Eduardo B. Fernandez and Rouyi Pan Presented by Liping Cai 03/15/2006.
SOUPA: Standard Ontology for Ubiquitous and Pervasive Applications Harry Chen, Filip Perich, Tim Finin, Anupam Joshi Department of Computer Science & Electrical.
Tim Finin University of Maryland, Baltimore County 29 January 2013 Joint work with Anupam Joshi, Laura Zavala and our students SRI Social Media Workshop.
Atomate It! End-user Context- Sensitive Automation using Heterogeneous Information Sources on the Web Max Van Kleek et el. MIT Presented by Sangkeun Lee,
UMBC iConnect Audumbar Chormale, Dr. A. Joshi, Dr. T. Finin, Dr. Z. Segall.
A service-oriented middleware for building context-aware services Center for E-Business Technology Seoul National University Seoul, Korea Tao Gu, Hung.
Hiding in the Mobile Crowd: Location Privacy through Collaboration.
Semantic Web and Policy Workshop Panel Contribution Norman M. Sadeh School of Computer Science Carnegie Mellon University Director, e-Supply Chain Management.
Semantics for Cybersecurity and Privacy Tim Finin, UMBC Joint work with Anupam Joshi, Karuna Joshi, Zareen Syed andmany UMBC graduate students
Streaming Knowledge Bases Onkar Walavalkar, Anupam Joshi Tim Finin and Yelena Yesha University of Maryland, Baltimore County 27 October 2008.
Illinois Security Lab Privacy Sensitive Location Information Systems in Smart Buildings Jodie P. Boyer, Kaijun Tan, Carl A. Gunter Midwest Security Workshop,
Coastal Atlas Interoperability - Ontologies (Advanced topics that we did not get to in detail) Luis Bermudez Stephanie Watson Marine Metadata Interoperability.
Page 1 Alliver™ Page 2 Scenario Users Contents Properties Contexts Tags Users Context Listener Set of contents Service Reasoner GPS Navigator.
MyActivity: A Cloud-Hosted Ontology-Based Framework for Human Activity Querying Amin BakhshandehAbkear Supervisor:
I.H. TorosluESSW Workshop Budapest May 20, A Semantic based Privacy Framework for Web Services Arif Tumer, Asuman Dogac, Hakki Toroslu Middle East.
Laura Zavala, Radhika Dharurkar, Pramod Jagtap, Tim Finin, Anupam Joshi and Amey Sane University of Maryland, Baltimore County AAAI Workshop on Activity.
User Profiling using Semantic Web Group members: Ashwin Somaiah Asha Stephen Charlie Sudharshan Reddy.
Tim Finin University of Maryland, Baltimore County UMBC HON 300 / CMSC491, Spring 2012 Joint work with Anupam Joshi, Laura Zavala, Radhika Dharurkar, Pramod.
Temporal Ontology Shervin Daneshpajouh ce.sharif.edu/~daneshpajouh.
Computational Policies in a Need to Share Environment Tim Finin University of Maryland, Baltimore County SemGrail workshop, Redmond WA, 21 June 2007.
THE SEMANTIC WEB By Conrad Williams. Contents  What is the Semantic Web?  Technologies  XML  RDF  OWL  Implementations  Social Networking  Scholarly.
1/14/ :59 PM1/14/ :59 PM1/14/ :59 PM Research overview Koen Victor, 12/2007.
An Ontology-based Approach to Context Modeling and Reasoning in Pervasive Computing Dejene Ejigu, Marian Scuturici, Lionel Brunie Laboratoire INSA de Lyon,
1 An infrastructure for context-awareness based on first order logic 송지수 ISI LAB.
NSF Cyber Trust Annual Principal Investigator Meeting September 2005 Newport Beach, California UMBC an Honors University in Maryland Trust and Security.
Diagrams. Typically, we view the static parts of a system using one of the four following diagrams. 1. Class diagram 2. Object diagram 3. Component diagram.
1 Web Services for Semantic Interoperability and Integration Tim Finin University of Maryland, Baltimore County Dagstuhl, 20 September 2004
Anupam Joshi University of Maryland, Baltimore County Joint work with Tim Finin and several students Computational/Declarative Policies.
International Workshop Jan 21– 24, 2012 Jacksonville, Fl USA Model-based Systems Engineering (MBSE) Initiative Ontology Action Team INCOSE MBSE Workshop.
OWL (Ontology Web Language and Applications) Maw-Sheng Horng Department of Mathematics and Information Education National Taipei University of Education.
A Context Framework for Ambient Intelligence
Database Systems: Design, Implementation, and Management Tenth Edition
Components.
Analysis Classes Unit 5.
FaceBlock: Semantic Context-Aware Privacy for Mobile Devices
Understanding Android Security
Chapter 1: Introduction to Systems Analysis and Design
WP2 INERTIA Distributed Multi-Agent Based Framework
Semantic Event-based Service Oriented Architecture
Introduction to Design Patterns
SPARQL SPARQL Protocol and RDF Query Language
Rob Gleasure IS3320 Developing and Using Management Information Systems Lecture 14: Data-Flow Diagrams 1 (Context-Level.
Wireless Sensor Network Architectures
Distribution and components
Understanding the OSI Reference Model
Video and Sensor Network Architecture and Displays
Privacy Preservation in Context-Aware Systems
Web Services for Semantic Interoperability and Integration
Analyzing and Securing Social Networks
Securing Home IoT Environments with Attribute-Based Access Control
LOD reference architecture
Chapter 1: Introduction to Systems Analysis and Design
Understanding Android Security
Chapter 1: Introduction to Systems Analysis and Design
Ponder policy toolkit Jovana Balkoski, Rashid Mijumbi
QoS Metadata Status 106th OGC Technical Committee Orléans, France
Presentation transcript:

Mobile, Collaborative and Context-Aware Systems Laura Zavala, Radhika Dharurkar, Pramod Jagtap, Tim Finin, Anupam Joshi and Amey Sane AAAI Workshop on Activity Context Representation 07 August 2011 http://ebiquity.umbc.edu/p/539/

Sharing context information Peer to peer communication Opportunistic Gossiping Fixed devices reposit, share, and summarize Peer to peer communication among co-located nearby wireless devices based on opportunistic gossiping is used for sharing place information. Fixed devices such as sensors and access points (APs) can be used to reposit, share, and even summarize statistically the place information overheard from passing-by mobile devices. Sharing context information

Current standard: location based applications Data Streaming Feeds NOSQL databases

Goals Prototype and evaluate mobile context-aware applications that deliver better user experiences Use smartphones to capture key elements of context Share context among devices Develop a semantic model of context Address privacy issues

Context / situation recognitio

Context / situation recognition Focus on individual activity and conceptual place recognition Using smartphones as sensors we use probabilistic models for context recognition noise, ambience light, accelerometer, Wifis, bluetooth, call stats, phone settings, user calendar In house data collection program was used to collect data to train discriminative classifiers to learn to recognize context 5 users, 1 month, logging TRUE activity and place attached to phone readings (noise, light, etc.) Naive Bayes, Decision Trees, Support Vector Machines, and Bagging+DecisionTrees

Context / situation recognition (process overview) Feature Vector Time, Noise level in db (avg, min, max), accel 3 axis (avg, min, max, magnitude, wifis, … Train Classifiers Decision Trees Naïve Bayes SVM

Evaluation Experiments Varying the level of granularity on the activity Motion, Stationary Work, Home, Outdoors, Other In Meeting, In class, Watching TV, Reading, Sleeping, etc. Two different schemes Individual: training and testing on one person’s data. Across users: training with one person’s data and testing it with other’s.

Results – Comparing classifiers Accuracy higher for decision tree classifiers Even better with bagging SVMs slightly below decision trees Weak performance of Naive Bayes

Results – Generalizing activities

Results – Testing across users

Results – Time and Location

Results – Decision tree output model

HMMs: Likely sequences of activity Activities are the states and sensor readings such as noise are the observations

HMMs A transition diagram for the data one user. Shows likely sequences of activities for the user

Use of Semantic Technologies Sharing context Collaborative context recognition

General Interaction Architecture Sensors on devices used for contextual clues Context KB on each device Context shared with neighboring devices Devices interact directly or through services on the Internet Privacy policies specify user preferences for release of information The Figure depicts a general interaction architecture for this type of systems. Sensors on devices are used to obtain clues about the local context of the user. We could use for example the user’s motion and mobility tracking and ambient conditions such as light and noise. The network component opportunistically gathers and disseminates local context information to neighboring fixed or mobile wireless devices. Its policy engine verifies the release policies to ensure release of information in accordance to the user preferences. Devices might interact directly or through services on the Internet. Inferences such as current activity can be drawn from the information collected by the sensors, the context information gathered, and additional resources (e.g., the user calendar and open geolocation KBs). The sensor’s raw data as well as the inferred context knowledge is stored in a local knowledge base on the device. Context-aware applications and network components may use this context knowledge to enhance their functionality. The locally inferred context knowledge can also be sent to context-aware services located on the Internet. General Interaction Architecture

Current Work Semantic model of context Context / situation recognition Ontology Population of local KB on the devices Integration with FOAF and GeoNames Context / situation recognition Use of the device sensors, status, and settings to recognize current context Individual activity and conceptual place recognition User Privacy Privacy policies for sharing and obfuscating contextual information Prototype using the user preferences and the context KB on the device

Semantic context modeling

The Place ontology: Semantic model of a person’s context Light-weight, upper level context ontology Centered around the notions of Users Conceptual places Activities Roles Time Conceptual places such as “At Work” and “At Home“ Activities occur at places and involves users under certain roles The Place ontology: Semantic model of a person’s context

Context KB on the devices A KB on the device which conforms to the ontology Links to FOAF and GeoNames Use of Geonames to assert further spatial knwledge in the KB <gn:Feature rdf:about="http://sws.geonames.org/4372143/"> <gn:name>UMBC</gn:name> <wgs84_pos:lat>39.25543</wgs84_pos:lat> <wgs84_pos:long>-76.71168</wgs84_pos:long> <wgs84_pos:alt>61</wgs84_pos:alt> <gn:parentFeature rdf:resource="http://sws.geonames.org/4347790/"/> Baltimore County <gn:parentCountry rdf:resource="http://sws.geonames.org/6252001/"/>  United States <gn:parentADM1 rdf:resource="http://sws.geonames.org/4361885/"/>  Maryland <gn:parentADM2 rdf:resource="http://sws.geonames.org/4347790/"/>  Baltimore County </gn:Feature> We use the Android Location API to obtain the position of the device. Position on Android phones is determined through location providers such as the device’s GPS and the network (which is based on availability of cell tower and WiFi access points). Given the Position of the user’s device, we assert the corresponding triples into the KB Context KB on the devices

Privacy preservation

Privacy Preservation Privacy controls in existing location sharing applications “Friends Only” and “Invisible” restrictions are the most prevalent Need for high-level, flexible, expressive, declarative policies Temporal restriction, freshness, granularity, access model (e.g. optimistic or pessimistic) Context dependent release of information Obfuscation of shared information Temporal Restriction: Date and time interval restrictions for disclosing the context information.  Freshness: Specifies the freshness of the disclosed context information. Timestamp: Specifies the time in which the privacy rule has been created.  AccessPolicy: Represents the access policy (Optimistic or Pessimistic) that this privacy rule is associated with. high-level, declarative policies that describe users’ information sharing preferences under given contextual situations. Besides being able to specify which information a user is willing to share, we can specify how that information should be shared. A user can disclose information with different accuracy levels; for instance, she may be willing to reveal to her close friends the exact room and building on which she is located, but only the vicinity or town to others.

Privacy Policies Requests come from other devices asking to share contextual information A specified protocol SPARQL queries Rules using context model and KB on device @prefix kb: <http://semantic-context.org/device#>. @prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>. @prefix foaf: <http://xmlns.com/foaf/0.1/> .   [AllowFamilyRule: (?requester kb:contextAccess kb:userPermitted) <- (?requester rdf:type kb:requester) (?groupFamily foaf:member ?requester) (?groupFamily foaf:name "Family") ] Prototype --Jena on Android-- Share building-wide location with teachers on weekdays only between 9 am and 6 pm Share detailed context information with family members Do not share my context if I am in a date with girlfriend Share my room-wide location with everyone in the same building as me Flexible privacy policies Role and group based context/location sharing Obfuscation of location nand activity information Summarization Current implementation of rules and ontology reasoning in Java using Jena A prototype has been implemented by one of our students using Jena semantic Web framework on Android phones. Requests are simple messages. Access rights are obtained by performing forward reasoning. Manual construction of answer. Ad-hoc location hierarchy provided for testing purposes.

Location Generalization Share my location with teachers on weekdays from 9am-5pm User’s exact location in terms of GPS co-ordinates is shared The user may not be interested to share GPS co-ordinates but fine with sharing city-level location Share my building-wide location with teachers on weekdays from 9am-5pm This approach has its own limitations as it doesn’t allow sharing on different granularity levels of the location. In many cases the user might be interested to share the location but not in terms of GPS coordinates. 9/22/2018

Location Generalization Hierarchical model of location to support location generalization The transitive “Part Of” property creates the location hierarchy GeoNames spatial containment knowledge is also used when populating the KB 9/22/2018

Activity Generalization Share my activity with friends on weekends User’s current activity is shared with friends on weekends share more generalized activity rather that precise confidential project meeting => Working, Date => Meeting User clearly needs to obfuscate certain pieces of activity information to protect her context information Share my public activity with friends on weekends Public is a visibility option It will enable users to have default privacy policies based on different accuracy levels. 9/22/2018

Activity Generalization

Privacy policies expressivity Policy language for policy declaration and enforcement integration Richer policies at the triple level Protect the inferences that can be drawn from the information that is shared A mix of rich pattern matching such as SPARQL and rules, with First Order semantics