Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Context and Discovery

Context and Discovery

Elvis D'Souza

April 26, 2013
Tweet

More Decks by Elvis D'Souza

Other Decks in Technology

Transcript

  1. @elvisds Outline ▪ Being  in  Context   ▪ Collecting  Information

      ▪ Delivering  Value   ▪ Challenges   ▪ Our  take  at  Pugmarks.me
  2. @elvisds Outline ▪ Being  in  Context   ▪ Collecting  Information

      ▪ Delivering  Value   ▪ Challenges   ▪ Our  take  at  Pugmarks.me
  3. @elvisds Outline ▪ Being  in  Context   ▪ Collecting  Information

      ▪ Delivering  Value   ▪ Challenges   ▪ Our  take  at  Pugmarks.me
  4. @elvisds Sensors  &  Wearable  Computing Mobile  Phone   – Camera

      – Microphone   – Social  Activity   – Touch   – Smell?
  5. @elvisds Sensors  &  Wearable  Computing ▪ Smart  Watches    

     Pebble,  Metawatch,  Basis,  iWatch?   ▪ Smart  Cars      Automatic,  Zipcar,  Waze   ▪ Smart  Calendars      Tempo  AI,  EasilyDo   ▪ Google  Glass
  6. @elvisds Outline ▪ Being  in  Context   ▪ Collecting  Information

      ▪ Delivering  Value   ▪ Challenges   ▪ Our  take  at  Pugmarks.me
  7. @elvisds Signals  available • Location   • WLAN   •

    Accelerometer   • Cell  Tower   • Call  Log   • SMS  Log   • Browsing  History   • Contacts   • Calendar   • Camera   • Running  Apps   • Installed  Apps
  8. @elvisds Signals  available  soon • Sleep  Patterns   • Pulse

      • Bio  Sensors   • Travel  Patterns   • Voice   • Data  from  Cars
  9. @elvisds Outline ▪ Being  in  Context   ▪ Collecting  Information

      ▪ Delivering  Value   ▪ Challenges   ▪ Our  take  at  Pugmarks.me
  10. @elvisds The  Context  Engine • Context  engines  have  to  be

     there  with  you,   when  you  are  doing  things.     • Context  engines  have  to  be  extremely   interesting  and  precise.   • Context  engines  are  personal,  and  should  offer   personalized  recommendations.   • Personal  assistants  are  long  term  companions.
  11. @elvisds The  Context  Engine You Your  likes   and  dislikes

    Your  goals   and  wishes Your  location What  are  you  going   to  do  soon? What  are  you  doing   right  now? Time Familiar?   New? Periodic? Planned? Normal? New? Your  actions People’s  actions Longer  Term Right  Now
  12. @elvisds Discovery  (Recommender  Systems) • Search  without  the  query  

    • Your  Web  v/s  Their  Web  v/s  Our  Web   • Recommendations  lead  to   – More  than  50%  of  Connections/Job  Applications/ Group  Joins  on  LinkedIn   – More  than  60%  of  clicks  on  YouTube  Homepage
  13. @elvisds Outline ▪ Being  in  Context   ▪ Collecting  Information

      ▪ Delivering  Value   ▪ Challenges   ▪ Our  take  at  Pugmarks.me
  14. @elvisds Information  Silos • Data  locked  up  in  products  

    – Social  Data  in  Facebook   – Information  Network  in  Twitter   – Professional  Network  in  LinkedIn
  15. @elvisds Privacy • Am  I  being  tracked?   • Will

     I  be  profiled  and  subjected  to   advertisements?   • Will  they  sell  my  data  to  3rd  parties?   • Who  is  responsible?
  16. @elvisds Privacy Who  knows  (a  lot)  about  you  already?  

    • Your  Car   • Your  Banker   • Your  Smartphone   • Your  Telecom  provider   • Your  Postman/Courier  service
  17. @elvisds Other  Challenges • Is  it  cool?  The  technology  &

     fashion   perspective   • Government  Regulations   • Adoption  by  people
  18. @elvisds Outline ▪ Being  in  Context   ▪ Collecting  Information

      ▪ Delivering  Value   ▪ Challenges   ▪ Our  take  at  Pugmarks.me
  19. @elvisds Pugmarks.me  –  Guiding  Principles ▪ Your  context  can  unearth

     valuable  information   ▪ Information  should  come  to  you  =  Discovery   ▪ The  product  must  not  have  to  chose  between   Users  and  Recommendations  as  the  product
  20. @elvisds Pugmarks.me  –  Guiding  Principles ▪ What  you  read  during

     your  day  (and  not  in  the   morning)  should  be  your  News   ▪ Discovery  must  be  serendipitous
  21. @elvisds Understanding  Context ▪ Is  this  an  Article  page?  A

     Hub  Page?  Are  you  on   Twitter?  On  Hacker  News?   ▪ Do  I  have  more  People  results  than  Article   results?
  22. @elvisds Privacy  Concerns? ▪ Most  people  trust  us  with  their

     data  or  do  not   care  (or  do  not  know?)   ▪ Quite  a  few  people  did  reach  out  to  us   ▪ What  we  do?   ▪ Store  activity  based  on  explicit  actions  by  users   ▪ Use  technology  to  avoid  storing  sensitive  data
  23. @elvisds Status  and  Road  Ahead ▪ Public  Beta:  600  users

     &  counting   ▪ Improve  ability  to  detect  context   ▪ Go  from  delivering  articles  to  delivering  deeper   insights   ▪ Go  Mobile