Presentation held by Zhong Wang and Zhimeng Chen at the Agricultural Ontology Service (AOS) Workshop 2012 in Kutching, Sarawak, Malaysia from September 3 - 4, 2012
Similar to Proposing a Semantic Multilingual Social Question Answering Service for Global Agricultural Practitioners: from Perspective of End Users (20)
Proposing a Semantic Multilingual Social Question Answering Service for Global Agricultural Practitioners: from Perspective of End Users
1. Proposing A Semantic Multilingual Social Question
Answering Service for Global Agricultural Practitioners:
from Perspective of End Users
Zhong WANG and Zhimeng CHEN
Sept 4, 2012 Kuching, Malaysia
2. Contents
Brief on Question Answering (QA) Services
QA for agricultural practitioners? - A case study
Draft proposing
Discussions and Conclusions
4. QA research and service
fact, list, definition, How , Why, hypothetical,
semantically constrained, and cross-lingual
Arbitrary question
Search
QA
service (un)structured
End user Answer Data
Relevant information
1. http://en.wikipedia.or g/wiki/Question_answering
5. Experienced services
keyword input (automatic)
search engine (SE) list of information
Knowledge portals list of fact
Natural language input (automatic/ human involved)
Text/syntactic analysis actual answer (TREC QA track data)
Entity/semantic analysis list of linked entities
Community/social people unpredicted result
6. Social QA service
a large number of real persons, entities, and relations/links
Exciting part:
Fast speed of information transfer
Dynamic increase and update of contents
Surprises from time to time
Not so exciting:
Unknown responding time
Unknown content quality
Maria Soledad Pera and Yiu-Kai Ng,
A Community Question-Answering Refinement System.
HT’11, June 6–9, 2011, Eindhoven, The Netherlands. The 22nd ACM Hypertext Conference
7. QA case in the Ag sector
Questions from a particular field farmer
in local spoken language
Insufficient/unclear descriptions
Impatient
Answers from the local service agent
piece of cake knows well
wait a moment not for sure New question
generated
does not know where to find it?
9. The Question
“where can I find some information on rice cultivation
in the dry area of Africa?”
Feedbacks against the NL input
Search Engine
Knowledge portal
SQA site
10. SE: Google
List of candidate links
Related may pumps
up on the top if lucky
Clicks into for …
19. Divergent feedbacks
General-purposed sites
not an exciting topic attract people
People answers for credits
90-9-1 Rule
http://www.useit.com/alertbox/participation_inequality.html
22. Resources required
People who know the question well, or
Domain relations know other entities that are easily located for
further help
Comprehensive but structured feedbacks relevant info still
valuable beside the actual answer
Worldwide connecting people speaking different language
23. We almost have them already
People Ag communities worldwide
Domain relations AGROVOC, Linked Ag Data, AGRIVIVO
Comprehensive but structured feedbacks OpenAgris,
AgriDrupal (w/ QA module), AgHarvest
Worldwide Statistical Machine Translation Services (SMTS)
24. (LOA: PO245395)
A Data Input System that Evaluates and Tests
Existing Technology of Machine Translation
from Chinese to English and Vise Versa
More on SMTS
25. What does it look like?
Africa China
Local AgIS QA Local AgIS QA
System System
End user (AgriDrupal) (AgriDrupal) End user
Local DB in Local DB in
French Chinese
(AgriVIVO)
Multilingual and Semantic Linked
Agricultural Dataset
26. What can we expect?
People in this domain
More convergent answers
Knowledge from colleagues world wide
Faster responses
29. Promotions and Sustainability
It is necessary to develop, deploy and evaluate
SQA service
The most direct way to make semantic techniques to
benefit the end users
User growth