Adrien Joly PhD Candidate Social Communications Dept. Alcatel-Lucent Bell Labs France SKP @ Villarceaux, France, 6/11/2009...
Abstract: context & scope <ul><li>PhD Subject: «  Contextual Filtering of Computer-Mediated Social Interactions  » </li></...
Outline <ul><li>Motivation </li></ul><ul><li>Approach </li></ul><ul><li>Implementation + Demo </li></ul><ul><li>Validation...
1 Motivation
Motivation  Approach  Implementation  Validation  Conclusion   Introduction to Awareness <ul><li>Awareness  is the state o...
Motivation  Approach  Implementation  Validation  Conclusion   Awareness is a natural ability in the real world <ul><ul><l...
Motivation  Approach  Implementation  Validation  Conclusion   Why does Awareness matter? <ul><li>Because: </li></ul><ul><...
Motivation  Approach  Implementation  Validation  Conclusion   Computer-supported communication tools for Awareness 1980 1...
Motivation  Approach  Implementation  Validation  Conclusion   Web « 2.0 » social / communication tools <ul><li>Social Net...
Motivation  Approach  Implementation  Validation  Conclusion   Web « 2.0 » social / communication tools <ul><li>Social Net...
Motivation  Approach  Implementation  Validation  Conclusion   Our proposal <ul><li>Social Networking Platforms  increase ...
2 Approach
Motivation  Approach  Implementation  Validation  Conclusion   Filtering possibilities <ul><li>Let’s build a social filter...
Motivation  Approach  Implementation  Validation  Conclusion   Similarity of context, our hypothesis <ul><ul><li>C A  is t...
Motivation  Approach  Implementation  Validation  Conclusion   Similarity of context, our hypothesis <ul><ul><li>C A  is t...
Motivation  Approach  Implementation  Validation  Conclusion   What is context ? <ul><li>Context [Dey, 2001]  : </li></ul>...
Motivation  Approach  Implementation  Validation  Conclusion   Requirements for sharing context <ul><ul><li>Context is pri...
Motivation  Approach  Implementation  Validation  Conclusion   Proposed approach illustrated <ul><li>Christine’s contextua...
3 First implementation
Motivation  Approach  Implementation  Validation  Conclusion   Towards a first implementation <ul><li>The proposed approac...
Motivation  Approach  Implementation  Validation  Conclusion   The Contextual Notification Framework - architecture Contex...
Motivation  Approach  Implementation  Validation  Conclusion   The Contextual Notification Framework – current implementat...
Motivation  Approach  Implementation  Validation  Conclusion   How to synthesize the contextual tag cloud from web browsin...
Motivation  Approach  Implementation  Validation  Conclusion   How to synthesize the contextual tag cloud from web browsin...
Motivation  Approach  Implementation  Validation  Conclusion   Contextual Tag Cloud algebra (work-in-progress) Sample tag ...
Motivation  Approach  Implementation  Validation  Conclusion   Context Aggregation and Filtering process Context aggregato...
4 Validation
Motivation  Approach  Implementation  Validation  Conclusion   To be validated <ul><ul><li>Hypothesis: </li></ul></ul><ul>...
Motivation  Approach  Implementation  Validation  Conclusion   Proposed trials (for validation) <ul><li>A) Online experime...
Motivation  Approach  Implementation  Validation  Conclusion   Proposed trials (for validation) <ul><li>B) Offline experim...
Conclusion
Motivation  Approach  Implementation  Validation  Conclusion   Conclusion <ul><li>Contributions: </li></ul><ul><ul><li>Sta...
www.alcatel-lucent.com www.alcatel-lucent.com Thank you for your attention! Please share your questions and comments! 
Upcoming SlideShare
Loading in...5
×

Tangible Contextual Tag Clouds towards Controlled and Relevant Social Interaction Awareness

1,838

Published on

Presented by Adrien Joly at Bell Labs France during a "SKP" session, this slideshow includes a motivated introduction to his phd thesis subject about contextual filtering of social interactions, its technical approach relying on "contextual tag clouds", and its current state of research.

Published in: Technology, Education
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
1,838
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide
  • Divider Section Break Pages Title block set 28/32pt Trebuchet white, flush left with text box positioned at, Horizontal 1.87” and Vertical 2.87” This slide is to be used to create Section Divider slides Please copy, paste and modify this slide to reflect your specific needs. It is not generated from a Master slide template
  • Divider Section Break Pages Title block set 28/32pt Trebuchet white, flush left with text box positioned at, Horizontal 1.87” and Vertical 2.87” This slide is to be used to create Section Divider slides Please copy, paste and modify this slide to reflect your specific needs. It is not generated from a Master slide template
  • Divider Section Break Pages Title block set 28/32pt Trebuchet white, flush left with text box positioned at, Horizontal 1.87” and Vertical 2.87” This slide is to be used to create Section Divider slides Please copy, paste and modify this slide to reflect your specific needs. It is not generated from a Master slide template
  • Divider Section Break Pages Title block set 28/32pt Trebuchet white, flush left with text box positioned at, Horizontal 1.87” and Vertical 2.87” This slide is to be used to create Section Divider slides Please copy, paste and modify this slide to reflect your specific needs. It is not generated from a Master slide template
  • Divider Section Break Pages Title block set 28/32pt Trebuchet white, flush left with text box positioned at, Horizontal 1.87” and Vertical 2.87” This slide is to be used to create Section Divider slides Please copy, paste and modify this slide to reflect your specific needs. It is not generated from a Master slide template
  • Transcript of "Tangible Contextual Tag Clouds towards Controlled and Relevant Social Interaction Awareness"

    1. 1. Adrien Joly PhD Candidate Social Communications Dept. Alcatel-Lucent Bell Labs France SKP @ Villarceaux, France, 6/11/2009 Tangible Contextual Tag Clouds towards Controlled and Relevant Social Interaction Awareness
    2. 2. Abstract: context & scope <ul><li>PhD Subject: «  Contextual Filtering of Computer-Mediated Social Interactions  » </li></ul><ul><li>Current project: Enterprise Contextual Notifier </li></ul><ul><li>Goal: Support effort-less and relevant exchange of fresh knowledge and statuses between people </li></ul><ul><li>Research focus: Context-Awareness applied to Social Networks </li></ul><ul><li>Publication : </li></ul><ul><ul><li>A. Joly, P. Maret, and J. Daigremont, &quot; Co ntext-awareness, the missing block of social networking,&quot; in 9th IFIP Working Conference on Virtual Enterprise, September 2008. </li></ul></ul>
    3. 3. Outline <ul><li>Motivation </li></ul><ul><li>Approach </li></ul><ul><li>Implementation + Demo </li></ul><ul><li>Validation </li></ul><ul><li>Conclusion + next steps </li></ul>
    4. 4. 1 Motivation
    5. 5. Motivation Approach Implementation Validation Conclusion Introduction to Awareness <ul><li>Awareness is the state or ability to perceive, to feel, or to be conscious of events, objects or sensory patterns […] without necessarily implying understanding. </li></ul><ul><li>wikipedia.org </li></ul><ul><li>Context Awareness </li></ul><ul><li>(location, surrounding environment…) </li></ul><ul><li>Peripheral Awareness </li></ul><ul><li>(current/recent people activities, moods, availability, status…) </li></ul>
    6. 6. Motivation Approach Implementation Validation Conclusion Awareness is a natural ability in the real world <ul><ul><li>Conversations, gossip and small talk </li></ul></ul><ul><ul><ul><li>Phone calls </li></ul></ul></ul><ul><ul><ul><li>Family dinner </li></ul></ul></ul><ul><ul><ul><li>Social events </li></ul></ul></ul><ul><ul><li>Awareness in the office </li></ul></ul><ul><ul><ul><li>Corporate / official communication </li></ul></ul></ul><ul><ul><ul><li>Seminar, conferences </li></ul></ul></ul><ul><ul><ul><li>Subscriptions (mailing lists etc…) </li></ul></ul></ul><ul><ul><ul><li>Team meetings </li></ul></ul></ul><ul><ul><ul><li>Watercooler / coffee machine breaks </li></ul></ul></ul><ul><ul><li>Disclaimer: « awareness » does not necessarily imply spying! </li></ul></ul>
    7. 7. Motivation Approach Implementation Validation Conclusion Why does Awareness matter? <ul><li>Because: </li></ul><ul><li>1) We care about people we know, people we relate to, people we meet </li></ul><ul><li>2) We are curious (eager to learn) </li></ul><ul><li>3) We don’t want to interrupt (i.e. availability) </li></ul><ul><li>4) We want to avoid clumsy, awkward situations </li></ul><ul><li>5) We want to position ourselves among others (i.e. competition) </li></ul><ul><li>(This list is not exhaustive) </li></ul>
    8. 8. Motivation Approach Implementation Validation Conclusion Computer-supported communication tools for Awareness 1980 1970 2000 Ambient Awareness ? 1990 Computer-Supported Cooperative Work Modems Bulletin Board Systems, Internet Relay Chat, Minitel, Internet Groupware / CSCW EIES (Electronic Information Exchange System) for Research, Professional communication and collaboration Perpetual proximity ? Communication becomes ubiquitous, instantaneous and seamless. Social web Forums, chat, opinions, user profiles, viral sharing, social networking sites 2010 Social Networking Sites IRC Minitel Web 1.0 Web 2.0 Internet Forums Cell telephony Mobile internet Ubiquitous computing Wireless networks Early networks ?
    9. 9. Motivation Approach Implementation Validation Conclusion Web « 2.0 » social / communication tools <ul><li>Social Networking Platforms increase Awareness </li></ul>
    10. 10. Motivation Approach Implementation Validation Conclusion Web « 2.0 » social / communication tools <ul><li>Social Networking Platforms increase Awareness </li></ul><ul><li>But it can steal a lot of attention  productivity loss </li></ul>INFORMATION OVERLOAD 2.0 NOW!!!
    11. 11. Motivation Approach Implementation Validation Conclusion Our proposal <ul><li>Social Networking Platforms increase Awareness </li></ul>Filter “ Aware” user Contributions Activities Status Updates Contacts Resources Needed <ul><li>Social updates </li></ul>and productive
    12. 12. 2 Approach
    13. 13. Motivation Approach Implementation Validation Conclusion Filtering possibilities <ul><li>Let’s build a social filter then! </li></ul><ul><li>But what filtering criteria should we adopt to get relevant updates ? </li></ul><ul><ul><li>Freshness ? (most recent updates) </li></ul></ul><ul><ul><li>Visibility ? (most viewed updates) </li></ul></ul><ul><ul><li>Popularity ? (most spread updates) </li></ul></ul><ul><ul><li>Response ? (most commented updates) </li></ul></ul><ul><ul><li>Similarity of context </li></ul></ul>
    14. 14. Motivation Approach Implementation Validation Conclusion Similarity of context, our hypothesis <ul><ul><li>C A is the context of a user U A sharing a piece of information I A . </li></ul></ul><ul><ul><li>C X is the context of a user U X that is a potential recipient of this information. </li></ul></ul>Hypothesis: I A is relevant to U X if C A is similar to C X A A = Travel in Asia U A = Alice I A = « Check out my amazing picture ! » A B = Working Java U B = Bob I B = « What database should I use ? » A C = Browsing map U C = Christine I C = « Looking for holiday locations… »
    15. 15. Motivation Approach Implementation Validation Conclusion Similarity of context, our hypothesis <ul><ul><li>C A is the context of a user U A sharing a piece of information I A . </li></ul></ul><ul><ul><li>C X is the context of a user U X that is a potential recipient of this information. </li></ul></ul>Hypothesis: I A is relevant to U X if C A is similar to C X C A = Travel, Asia C C = Travel C B = Java Dev. A A = Travel in Asia U A = Alice I A = « Check out my amazing picture ! » A B = Working Java U B = Bob I B = « What database should I use ? » A C = Browsing map U C = Christine I C = « Looking for holiday locations… » Similar context: travel No relevant match for this context
    16. 16. Motivation Approach Implementation Validation Conclusion What is context ? <ul><li>Context [Dey, 2001] : </li></ul><ul><li>«  any information that can be used to characterize the situation of an entity  » </li></ul><ul><ul><li>From physical sensors: </li></ul></ul><ul><ul><ul><li>GPS location </li></ul></ul></ul><ul><ul><ul><li>Surrounding bluetooth devices and people </li></ul></ul></ul><ul><ul><ul><li>Temperature ? Not really useful to increase relevance, but why not! </li></ul></ul></ul><ul><ul><li>From computer-mediated actions: </li></ul></ul><ul><ul><ul><li>Current and recent communication activity (emails, SMS, phone calls…) </li></ul></ul></ul><ul><ul><ul><li>Web browsing activity => topic(s) of currently browsed web pages </li></ul></ul></ul><ul><ul><ul><li>Analysis of user-generated statuses ? </li></ul></ul></ul>
    17. 17. Motivation Approach Implementation Validation Conclusion Requirements for sharing context <ul><ul><li>Context is private  how to control privacy ? </li></ul></ul><ul><ul><ul><li>Make this context « tangible »: </li></ul></ul></ul><ul><ul><ul><li>Easy to visualize (to see quickly what information is involved) </li></ul></ul></ul><ul><ul><ul><li>Easy to manipulate (to select what will be visible to the system) </li></ul></ul></ul><ul><ul><ul><li>Easy to share, on demand only </li></ul></ul></ul><ul><ul><li>Context from heterogeneous sources  what model to adopt ? </li></ul></ul><ul><ul><ul><li>Ontologies / semantic technologies ? (closed-world, too static, and costly) </li></ul></ul></ul><ul><ul><ul><li>Folksonomies: descriptions of contextual resources entered by people </li></ul></ul></ul><ul><ul><li>Context is fuzzy and covers various granularities </li></ul></ul><ul><ul><ul><li>The « tag cloud » visualization is a good candidate to display weighted keywords and navigate through them </li></ul></ul></ul>
    18. 18. Motivation Approach Implementation Validation Conclusion Proposed approach illustrated <ul><li>Christine’s contextual cloud: </li></ul><ul><li>McDonalds Chatelet DriveIn Radio Alice France Californication Paris RedHotChiliPeppers Wifi_SSID_5874 </li></ul>Mario is shopping near Chatelet Mario: «  let’s grab a coffee at SB !  » Alice has just shared a photo [ view ] Lucie is listening to Californication [ i like this ] Kevin: «  new McChicken is great!  » Car incident 32 meters away Alice GPS Wifi McDonald’s restaurant Radio currently playing a song Christine
    19. 19. 3 First implementation
    20. 20. Motivation Approach Implementation Validation Conclusion Towards a first implementation <ul><li>The proposed approach relies on a « smart » environment, hard to simulate… </li></ul><ul><li> let’s consider a computer-mediated environment first: the enterprise </li></ul><ul><li>Context that can be gathered from a user’s computer: </li></ul><ul><ul><li>Web sites being browsed </li></ul></ul><ul><ul><li>Communication sessions </li></ul></ul><ul><ul><li>Documents being worked on </li></ul></ul><ul><ul><li>Applications being used </li></ul></ul><ul><li>Publication : </li></ul><ul><ul><li>A. Joly, P. Maret, and J. Daigremont, “Workspace Awareness without Overload: Contextual Filtering of Social Interactions”. In Smart Offices and Other Workspaces, workshop of the Intelligent Environments 2009 conference, Barcelona, Spain. 2009. </li></ul></ul>
    21. 21. Motivation Approach Implementation Validation Conclusion The Contextual Notification Framework - architecture Context aggregator Desktop activity sniffers Social notifier Contextual Filter Enterprise worker User’s workstation Infrastructure Interaction logs (REST interface) High level context Proposed notifications
    22. 22. Motivation Approach Implementation Validation Conclusion The Contextual Notification Framework – current implementation Context aggregator Desktop activity sniffers Social notifier Contextual Filter Enterprise worker User’s workstation Infrastructure Web navigation and sharing events (open, focus, close) Contextual cloud Proposed notifications Lightweight JAVA RESTful server Firefox plug-in Jetty-based JAVA server Persevere JSON DB
    23. 23. Motivation Approach Implementation Validation Conclusion How to synthesize the contextual tag cloud from web browsing ? The user opens a web page…
    24. 24. Motivation Approach Implementation Validation Conclusion How to synthesize the contextual tag cloud from web browsing ? Expressivity Low level and static author description Automatic content analysis Mining semantic concepts from content People-entered tags (wisdom of crowds) 1) URL is sent to the Context Aggregator 2) Content is analyzed by enhancers (including web services)
    25. 25. Motivation Approach Implementation Validation Conclusion Contextual Tag Cloud algebra (work-in-progress) Sample tag cloud R: (normalized) <ul><li>Research report : </li></ul><ul><ul><li>A. Joly, P. Maret, “Contextual Tag Cloud Algebra”, ALU / LIRIS </li></ul></ul>0.1 0.1 0.3 0.5 « Discount » « Flight » « Asia » « Travel »
    26. 26. Motivation Approach Implementation Validation Conclusion Context Aggregation and Filtering process Context aggregator Firefox plug-in Contextual Filter User’s workstation Infrastructure U: u i = URLs R, S = Contextual tag clouds Relevant updates filtered by contextual similarity Enhancer 1 Enhancer 2
    27. 27. 4 Validation
    28. 28. Motivation Approach Implementation Validation Conclusion To be validated <ul><ul><li>Hypothesis: </li></ul></ul><ul><ul><li>Recommended social updates are relevant and interesting (algorithms) </li></ul></ul><ul><ul><li>Users will share their contextual cloud frequently (privacy & ergonomics) </li></ul></ul><ul><ul><li>Users will post social updates (usage) or syndicate their existing social feeds </li></ul></ul><ul><li>Algorithms: </li></ul><ul><ul><li>Tag cloud synthesis  well represented contexts ? </li></ul></ul><ul><ul><li>Tag cloud similarity  sufficient recall and precision ? </li></ul></ul>
    29. 29. Motivation Approach Implementation Validation Conclusion Proposed trials (for validation) <ul><li>A) Online experimentation: application usage </li></ul><ul><ul><li>Requirements: 10+ Firefox users in the lab, duration: 3+ weeks </li></ul></ul><ul><ul><li>Validate the robustness and scalability of the filtering server </li></ul></ul><ul><ul><li>Provide and deploy Firefox plug-in and context aggregator </li></ul></ul><ul><ul><li>Train the users and gather feedback all along the trial period </li></ul></ul><ul><li>(+) Feedback from actual application usage can be gathered </li></ul><ul><li>(-) The algorithms must be stable </li></ul><ul><li>(-) High investment expected from trial participants </li></ul><ul><li>Risk: low usage  no sufficient feedback for validation </li></ul>
    30. 30. Motivation Approach Implementation Validation Conclusion Proposed trials (for validation) <ul><li>B) Offline experimentation: data gathering </li></ul><ul><ul><li>Requirements: 5+ Firefox users (with social feeds) in the lab, duration: 1 week </li></ul></ul><ul><ul><li>Provide a transparent (or not?) Firefox plug-in to log web activity </li></ul></ul><ul><ul><li>Contextual clouds and recommendations can be generated offline </li></ul></ul><ul><li>(+) Results from several algorithms can be evaluated by users at once </li></ul><ul><li>(-) No actual usage of the application  no feedback on application </li></ul>
    31. 31. Conclusion
    32. 32. Motivation Approach Implementation Validation Conclusion Conclusion <ul><li>Contributions: </li></ul><ul><ul><li>State-of-the-Art in: Context-Awareness (models), Computer-Mediated Collaborative Work (filtering techniques), Awareness systems (HCI) </li></ul></ul><ul><ul><li>A novel context modeling (and algebra) approach based on tag clouds </li></ul></ul><ul><ul><li>Context aggregation and filtering techniques to be tested </li></ul></ul><ul><ul><li>An innovative social application that maintains privacy control </li></ul></ul><ul><li>Next steps: </li></ul><ul><ul><li>Prepare and run a trial for validation of the approach </li></ul></ul><ul><ul><li>Publish experimental results </li></ul></ul><ul><ul><li>Generalize the approach with sensor-based context (using Social Bookmarker) </li></ul></ul>
    33. 33. www.alcatel-lucent.com www.alcatel-lucent.com Thank you for your attention! Please share your questions and comments! 

    ×