FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.
-
Upload
madison-veronica-bradford -
Category
Documents
-
view
219 -
download
1
Transcript of FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.
![Page 1: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/1.jpg)
FindAll: A Local Search Engine for Mobile Phones
Aruna BalasubramanianUniversity of Washington
![Page 2: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/2.jpg)
Co-AuthorsInformation Retrieval
Systems
Niranjan Balasubramanian, UW
Sam Huston, UMassDon Metzler, USC
(now Google)
David Wetherall, UW
![Page 3: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/3.jpg)
Mobile web search performance poor
• Order of magnitude slower on cellular networks
![Page 4: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/4.jpg)
Cellular connectivity is poorPew, April 2012
This work: Can we trade storage for connectivity to improve search performance?
![Page 5: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/5.jpg)
Leveraging re-finding.• Searching for a previously viewed page.
Mobile: 70% of searches for 50% users.Non-Mobile: 40% to 60% of all searches.
![Page 6: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/6.jpg)
FindAll local search engine• Search interface to search any previously viewed
page, on any of your device
![Page 7: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/7.jpg)
Is this the same as caching/history?• It is a search interface on top of caching: History
seldom used
• Is this same as Google history or chrome sync?
![Page 8: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/8.jpg)
What is a search interface?
• Uses indexes and retrieval algorithms for effective search– Keyword matching is easy but not effective– Database of search queries miss query changes
and non-searched web pages
Challenge: Search engines are memory/energy intensive
![Page 9: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/9.jpg)
Talk outline
• User study– Identifies re-finding behavior
• FindAll– Design of search engine for phones
• Evaluation– Results of tradeoffs in practice
![Page 10: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/10.jpg)
Talk outline
• User study– Identify re-finding behavior
• FindAll– Design of search engine for phones
• Evaluation– Results of tradeoffs in practice
![Page 11: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/11.jpg)
IR-approved study
• Monitored 23 participants for 1 month– Grad and under-grad students
• Collected logs from user’s mobile/desktop– Visited URL and search query (anonymized)
• Mark URL re-found if– Page revisited via search query, and unchanged
![Page 12: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/12.jpg)
Examples
![Page 13: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/13.jpg)
Re-finding accounts for 52% of search
Cross-device re-finding is 70%
>20% of re-finds have different query
Lots of opportunities to search locally.
![Page 14: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/14.jpg)
45% re-finding occurs within 50 minutes
Time between first visit and subsequent re-finding
Need to index when the page is first accessed.
![Page 15: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/15.jpg)
User’s show diverse re-finding patterns
Need to adapt to user
![Page 16: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/16.jpg)
User’s re-finding fairly constant
This user: Avg re-finding 43%, std deviation 9%
![Page 17: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/17.jpg)
User study summary
• Lots of opportunities to leverage re-finding
• Need to index near when page is accessed
• Need to adapt to users
![Page 18: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/18.jpg)
Talk outline
• User study– Identifies re-finding behavior
• FindAll– Design of search engine for phones
• Evaluation– Results of tradeoffs in practice
![Page 19: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/19.jpg)
FindAll architecture
Storage
Partial Indexes
Cache
![Page 20: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/20.jpg)
When to index?
High availability
High index energy
Low availability
Low index energy
![Page 21: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/21.jpg)
FindAll indexing
• Maximize availability, such that total energy consumption is no more than default search
Expected energy for indexing <=Expected energy if indexing not done (default
search)
FindAll estimates expectations based on user behavior
![Page 22: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/22.jpg)
Predicting user re-finding probability
• Online classier: What is the probability of a web page being re-found in the next T minutes.
• Classifier features1. base re-finding probability of user?2. user in a browsing session?3. web page been re-found recently?
![Page 23: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/23.jpg)
Prototype on Android • Adapt Galago search engine for phones– Implement partial indexing and merging
• Implement online energy cost estimator– Train classifier when mobile is charging– Make an indexing decision every 5 mins
![Page 24: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/24.jpg)
Talk outline
• User study– Identify re-finding behavior
• FindAll– Design of search engine for phones
• Evaluation– Results of tradeoffs in practice
![Page 25: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/25.jpg)
Evaluation goals• Benefits and Costs• Latency, Availability, 3G data usage• Energy, Storage
• Alternate approaches• Keyword, Database
• Alternate indexing strategies• Cloud index, Always index, Fixed index
Results based on prototype and user traces
![Page 26: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/26.jpg)
Evaluation goals• Benefits and Costs• Latency, Availability, 3G data usage• Energy, Storage
• Alternate approaches• Keyword, Database
• Alternate indexing strategies• Cloud index, Always index, Fixed index
![Page 27: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/27.jpg)
FindAll improves web page latency
3.42
1.82
![Page 28: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/28.jpg)
FindAll does not increase energy
![Page 29: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/29.jpg)
Availability under limited connectivity
43%
(Under a random 50% connectivity model)
![Page 30: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/30.jpg)
FindAll indexing important for energy benefits
![Page 31: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/31.jpg)
Conclusions
FindAll makes a win-win tradeoff for search– Decrease latency and increase availability, with
reduced energy and bandwidth
Future directionsSearch primitive: Integrating re-finding with other mobile apps
Context-based re-finding: Adding sensor cues to pages
![Page 33: FindAll: A Local Search Engine for Mobile Phones Aruna Balasubramanian University of Washington.](https://reader035.fdocuments.in/reader035/viewer/2022062713/56649ccb5503460f949939c4/html5/thumbnails/33.jpg)
Other results
• Static Indexing strategies – Increase energy by up to 50% compared to default
search for low re-find users– Decreases availability by up to 39% for high re-find
users
• Storage requirement less than 1.7GB per month