Transforming Personal Artifacts into Probabilistic Narratives Setareh Rafatirad and Kathryn Laskey 1 Setareh Rafatirad,

Slides:



Advertisements
Similar presentations
Dr. Leo Obrst MITRE Information Semantics Information Discovery & Understanding Command & Control Center February 6, 2014February 6, 2014February 6, 2014.
Advertisements

International Technology Alliance In Network & Information Sciences International Technology Alliance In Network & Information Sciences Paul Smart, Ali.
Privacy By Design Sample Use Case
Location-Based Social Networks Yu Zheng and Xing Xie Microsoft Research Asia Chapter 8 and 9 of the book Computing with Spatial Trajectories.
A Stepwise Modeling Approach for Individual Media Semantics Annett Mitschick, Klaus Meißner TU Dresden, Department of Computer Science, Multimedia Technology.
Evaluating Color Descriptors for Object and Scene Recognition Koen E.A. van de Sande, Student Member, IEEE, Theo Gevers, Member, IEEE, and Cees G.M. Snoek,
Outline What is a data warehouse? A multi-dimensional data model Data warehouse architecture Data warehouse implementation Further development of data.
1 A Description Logic with Concrete Domains CS848 presentation Presenter: Yongjuan Zou.
Contextual Augmentation of Ontology for Recognizing Sub-Events Setareh Rafatirad and Ramesh Jain ICSC2011, Sep 19-21,
Bring Order to Your Photos: Event-Driven Classification of Flickr Images Based on Social Knowledge Date: 2011/11/21 Source: Claudiu S. Firan (CIKM’10)
1 Overview of Image Retrieval Hui-Ying Wang. 2/42 Reference Smeulders, A. W., Worring, M., Santini, S., Gupta, A.,, and Jain, R “Content-based.
Gene selection using Random Voronoi Ensembles Stefano Rovetta Department of Computer and Information Sciences, University of Genoa, Italy Francesco masulli.
LEDIR : An Unsupervised Algorithm for Learning Directionality of Inference Rules Advisor: Hsin-His Chen Reporter: Chi-Hsin Yu Date: From EMNLP.
An Approach to Evaluate Data Trustworthiness Based on Data Provenance Department of Computer Science Purdue University.
PR-OWL: A Framework for Probabilistic Ontologies by Paulo C. G. COSTA, Kathryn B. LASKEY George Mason University presented by Thomas Packer 1PR-OWL.
1 © Ramesh Jain Social Life Networks: Ontology-based Recognition Ramesh Jain Contact:
Report on Intrusion Detection and Data Fusion By Ganesh Godavari.
Image Search Presented by: Samantha Mahindrakar Diti Gandhi.
Data Quality Class 3. Goals Dimensions of Data Quality Enterprise Reference Data Data Parsing.
Knowledge Acquisitioning. Definition The transfer and transformation of potential problem solving expertise from some knowledge source to a program.
Gimme’ The Context: Context- driven Automatic Semantic Annotation with CPANKOW Philipp Cimiano et al.
Generating Summaries and Visualization for Large Collections of Geo-referenced Photographs Alexander Jaffe*, Mor Naaman*, Tamir Tassa †, Marc Davis $ *Yahoo!
1 Human Detection under Partial Occlusions using Markov Logic Networks Raghuraman Gopalan and William Schwartz Center for Automation Research University.
Image Metadata Summary of 4/18/99 NISO/DLF Image Metadata Meeting ( Howard Besser UCLA School of Education & Information.
Software Process and Product Metrics
EVENT IDENTIFICATION IN SOCIAL MEDIA Hila Becker, Luis Gravano Mor Naaman Columbia University Rutgers University.
Predicting Missing Provenance Using Semantic Associations in Reservoir Engineering Jing Zhao University of Southern California Sep 19 th,
Amarnath Gupta Univ. of California San Diego. An Abstract Question There is no concrete answer …but …
OMAP: An Implemented Framework for Automatically Aligning OWL Ontologies SWAP, December, 2005 Raphaël Troncy, Umberto Straccia ISTI-CNR
An Intelligent Broker Architecture for Context-Aware Systems A PhD. Dissertation Proposal in Computer Science at the University of Maryland Baltimore County.
Computer vision.
The Data Attribution Abdul Saboor PhD Research Student Model Base Development and Software Quality Assurance Research Group Freie.
Watch, Listen and Learn Sonal Gupta, Joohyun Kim, Kristen Grauman and Raymond Mooney -Pratiksha Shah.
Information Systems & Semantic Web University of Koblenz ▪ Landau, Germany Semantic Web - Multimedia Annotation – Steffen Staab
Dart: A Meta-Level Object-Oriented Framework for Task-Specific Behavior Modeling by Domain Experts R. Razavi et al..OOPSLA Workshop DSML‘ Dart:
Interactive Discovery and Semantic Labeling of Patterns in Spatial Data Thomas Funkhouser, Adam Finkelstein, David Blei, and Christiane Fellbaum Princeton.
School of Computing FACULTY OF ENGINEERING Developing a methodology for building small scale domain ontologies: HISO case study Ilaria Corda PhD student.
Finding Better Answers in Video Using Pseudo Relevance Feedback Informedia Project Carnegie Mellon University Carnegie Mellon Question Answering from Errorful.
Workshop on Social Events in Web Multimedia, ICMR 2014 Social Event Detection at MediaEval: a three-year retrospect of tasks and results Georgios Petkos,
Report on Intrusion Detection and Data Fusion By Ganesh Godavari.
Digital Camera Basics Presented by Stephen Schneider June, 2013.
Dimitrios Skoutas Alkis Simitsis
Collective Vision: Using Extremely Large Photograph Collections Mark Lenz CameraNet Seminar University of Wisconsin – Madison February 2, 2010 Acknowledgments:
Module networks Sushmita Roy BMI/CS 576 Nov 18 th & 20th, 2014.
1 Transparent Metadata Capture for Environmental Science Wade Sheldon Georgia Coastal Ecosystems LTER University of Georgia
Automatic Video Tagging using Content Redundancy Stefan Siersdorfer 1, Jose San Pedro 2, Mark Sanderson 2 1 L3S Research Center, Germany 2 University of.
Dictionary based interchanges for iSURF -An Interoperability Service Utility for Collaborative Supply Chain Planning across Multiple Domains David Webber.
Chapter 10. The Explorer System in Cognitive Systems, Christensen et al. Course: Robots Learning from Humans On, Kyoung-Woon Biointelligence Laboratory.
Total Recall: Automatic Query Expansion with a Generative Feature Model for Object Retrieval O. Chum, et al. Presented by Brandon Smith Computer Vision.
Final Project Mei-Chen Yeh May 15, General In-class presentation – June 12 and June 19, 2012 – 15 minutes, in English 30% of the overall grade In-class.
An Ontology-based Approach to Context Modeling and Reasoning in Pervasive Computing Dejene Ejigu, Marian Scuturici, Lionel Brunie Laboratoire INSA de Lyon,
Instance Discovery and Schema Matching With Applications to Biological Deep Web Data Integration Tantan Liu, Fan Wang, Gagan Agrawal {liut, wangfa,
Department of Mathematics Computer and Information Science1 CS 351: Database Management Systems Christopher I. G. Lanclos Chapter 4.
Location-based Social Networks 6/11/20161 CENG 770.
Semantic Web Technologies Readings discussion Research presentations Projects & Papers discussions.
CPSC 322, Lecture 22Slide 1 Logic: Domain Modeling /Proofs + Top-Down Proofs Computer Science cpsc322, Lecture 22 (Textbook Chpt 5.2) Oct, 30, 2013.
Digital Camera Basics Presented by Stephen Schneider June, 2013.
Datamining : Refers to extracting or mining knowledge from large amounts of data Applications : Market Analysis Fraud Detection Customer Retention Production.
EXIF Data and Imagery Forensics
Ontology Evolution: A Methodological Overview
Web IR: Recent Trends; Future of Web Search
Data Integration with Dependent Sources
Lecture 12: Data Wrangling
Data Warehousing Data Mining Privacy
BPaaS Evaluation Environment Research Prototype
Image Metadata Summary of 4/18/99 NISO/DLF Image Metadata Meeting
A framework for ontology Learning FROM Big Data
Visual Grounding.
CoXML: A Cooperative XML Query Answering System
Presentation transcript:

Transforming Personal Artifacts into Probabilistic Narratives Setareh Rafatirad and Kathryn Laskey 1 Setareh Rafatirad, Kathryn Laskey, George Mason University (UAIW2013)

Outline Motivation Agglomerative Clustering Event Ontology Augmentation Filtering Evaluation Summary 2 Setareh Rafatirad, Kathryn Laskey, George Mason University

Motivation 3 Date/Time Original : 2009:12:15 11:46:44 Create Date : 2009:12:15 11:46:44 Shutter Speed Value : 1/304 Aperture Value : 2.6 Brightness Value : 7.16 GPS Version ID : Compression : JPEG (old-style) Thumbnail Offset : 1280 Thumbnail Length : 9508 Bits Per Sample : 8 Color Components : 3 Y Cb Cr Sub Sampling : YCbCr4:2:2 (2 1) Aperture : 2.6 GPS Altitude : 0 m Above Sea Level GPS Latitude : GPS Longitude : Shutter Speed : 1/304 Focal Length : 3.8 mm Light Value : 12.0 EXIF TAG Setareh Rafatirad, Kathryn Laskey, George Mason University

Motivation cont’d Expressive event tag – Multi-granular Conceptual description Containment event relationships e.g. subevent, during, etc. – Multi-adaptation of Contextual description Visit landmark Forbidden City in a trip to Beijing, visit Landmark Washington monument in Washington, DC. 4 Setareh Rafatirad, Kathryn Laskey, George Mason University

Motivation cont’d 5 Ontological Event models Data sources + Annotation technique Geo-tagged photo stream of an event + photo stream annotated with context-adaptive event ontology (probabilistic narratives) Setareh Rafatirad, Kathryn Laskey, George Mason University

Domain Event Ontology 6 Setareh Rafatirad, Kathryn Laskey, George Mason University

Perdurant Endurant Participant Spatial Region Interval occurs-during Literal:Timestamp start end occurs-at point double:lat double:lng hasLatitude hasLongitude Visual Concept visual-constraint Subevent containment Rules: If subevent(B,A), then: B.Start>= A.start && B.end<= A.end Contained-in(B.located-at,A.located-at) B.media ⊂ A.media B.participant ⊂ A.participant Subevent containment Rules: If subevent(B,A), then: B.Start>= A.start && B.end<= A.end Contained-in(B.located-at,A.located-at) B.media ⊂ A.media B.participant ⊂ A.participant subevent-of Trel Core Event Ontology 7 Setareh Rafatirad, Kathryn Laskey, George Mason University

Solution Strategy 8 Setareh Rafatirad, Kathryn Laskey, George Mason University

Challenges How to obtain expressive event tags? How to determine the event category? What kind of data sources should be used to compute the tags? 9 Setareh Rafatirad, Kathryn Laskey, George Mason University

Agglomerative Clustering 10 Setareh Rafatirad, Kathryn Laskey, George Mason University

Event Ontology Augmentation Definition1: – A context-adaptive event ontology is an instance event ontology, augmented with concrete context cues from disparate sources. Definition2: – A tag t for a group of photos C is an augmented instance of a subevent of event E that either exists in event ontology O, or can be derived from O such that t is the finest subevent that can be assigned to C. 11 Setareh Rafatirad, Kathryn Laskey, George Mason University

Event Ontology Augmentation cont’d Given a photo pj, find the sound cluster C containing pj Represent C with a set of consistent descriptors – using the descriptors of every pi C, – guided by the descriptors of pj Confidence of cluster descriptor d: 12 Setareh Rafatirad, Kathryn Laskey, George Mason University

Event Ontology Augmentation cont’d Context Discovery – Schema for source representation – SPARQL for query sources 13 SELECT ?var 1 FROM WHERE{ attr 1 class w. attr 2 class f. attr 3 class u. ?x rel a ?var 1. ?x rel b ?y. ?x rel c ?z. ?y rel d attr 1. ?z rel h attr 2. } weather StatisticalSource input_attr: (loc,t, zone); output_attr: (weather); loc Point; t Timestamp; zone TimeZone; Point Space; Point Literal:numeric; Point Literal:numeric. Timestamp Time; weather Ambiance; Ambiance Literal:String; Ambiance Quality. Setareh Rafatirad, Kathryn Laskey, George Mason University

Event Ontology Augmentation cont’d Descriptors consistency – Example 14 outdoorSeating : true; sceneT ype : outdoor; weatherCondition : storm Rule1: Rule2 is entailed: inconsistency detected! Setareh Rafatirad, Kathryn Laskey, George Mason University

Event Ontology Augmentation cont’d Event Inference – Find event categories – Rank event candidates through Measure of Plausibility Granularity score for an event candidate Context-Plausibility score for an event candidate Compare event candidates – Instantiate and augment the most plausible event candidate 15 Number of event constraints Score related to an event constraint Setareh Rafatirad, Kathryn Laskey, George Mason University

Context-Adaptive Event Ontology (Probabilistic Narratives) 16 Setareh Rafatirad, Kathryn Laskey, George Mason University

Filtering 17 Setareh Rafatirad, Kathryn Laskey, George Mason University

Experiments and Evaluation Formative evaluation 3 domain models 1M photos, 50 Albums from lab and Flickr Multiple Data Sources – Trip Advisor – Google Geocoding – Yelp – Upcoming – Evite – Facebook – Wunderground – Foursquare – Face.com – Pictorria (MIT SUN and YELP training set, 500 images/concept, 58 visual concepts, pyramids of color histogram and GIST features-Oliva et al.(2001), Hejrati et al.(2012)) – GoogleMovieShowTimes – GeoPlanet – Disneyland.disney.go.com Evaluation metrics – Average correctness – Average Context 18 Setareh Rafatirad, Kathryn Laskey, George Mason University

Experiments and Evaluation 19 Setareh Rafatirad, Kathryn Laskey, George Mason University

Experiments and Evaluation 20 Domain relevancy Setareh Rafatirad, Kathryn Laskey, George Mason University

Experiments and Evaluation 21 Setareh Rafatirad, Kathryn Laskey, George Mason University

Summary Improving performance in terms of quality of tags Evaluation measure Event ontology augmentation and information integration – Automated context discovery – Relaxation Policies – Validation using external sources – Plausibility Measure 22 Setareh Rafatirad, Kathryn Laskey, George Mason University

23 Setareh Rafatirad, Kathryn Laskey, George Mason University

Back up slides 24 Setareh Rafatirad, Kathryn Laskey, George Mason University

Related Work Event-Centric Models – Francois et al.(2005),Town at al.(2006), Neumann et al.(2008), Mezaris et al.(2010), Scherp et al.(2009), Gupta and Jain(2011), Masolo et al.(2002), Lagoze et al(2010). Joint-Context Event-Models – Viana et al.(2007,2008), Liu et al.(2011), Fialho et al.(2010), Cao et al. (2008), Paniagua et al.(2012). 25 Setareh Rafatirad, Kathryn Laskey, George Mason University

Event Ontology Augmentation cont’d Instantiation and augmentation/refinement – Iteration 1 26 TA l2 l1 WP GoldenGate Alcatraz Island hasName Setareh Rafatirad, Kathryn Laskey, George Mason University

Event Ontology Augmentation cont’d Instantiation and augmentation/refinement – Iteration 1 27 TA l2 l1 WP hasName hasCategory Alcatraz Island hasName hasCategory Prison, Historic site … … GoldenGate Toll Bridge, Historic Site Setareh Rafatirad, Kathryn Laskey, George Mason University

My Trip l1 l2 att 1,…,att n Perdurant Trip LunchShoppingvisitLandmark subevent-of subClass-of Spatial Region occurs-at Visit-1 Visit-2 occurs-at subevent-of Event Ontology Augmentation cont’d Verification 28 Setareh Rafatirad, Kathryn Laskey, George Mason University