40 likes | 161 Views
Semantic Context Service Delivery. Ideas by Sang- keun Lee IDS Lab, SNU. UI 1. * Semantic Search. * Recommendation System. Alg 1. content. …. Source 1. user. …. context. UI 1. Inference Engine. Source 1. Need to be combined. Log Manager. DB. …. Alg 1. Source n. DB.
E N D
Semantic Context Service Delivery Ideas by Sang-keun Lee IDS Lab, SNU
UI1 * Semantic Search * Recommendation System Alg 1 content … Source 1 user … context UI1 Inference Engine Source 1 Need to be combined Log Manager DB … Alg 1 Source n DB * Crawling Crawler DB
Service 1 Service 1 Service 1 Service n Query Manager Result Processing UI1 UI2 UI3 UIn Parser … Integration Interface Query Expansion Service Combination Module Query Request content … Query Mediation Source 1 user … Various Actions context Data Mediator Source 1 Log Mediator Push Service Request … Context Monitoring Matcher Log Manager Matcher decides which algorithm to be used Log Monitoring Data Monitoring Source n Service Rule Generator … Alg 1 Alg 2 Alg 3 DB & Index Service Rules / Logs We can see heterogeneous databases(physical) as an integrated ontology(logical) Ontology Manager Mediator Index Manager Inference Engine Crawler Synchronizer DB DB DB DB Configuration files define how to manage each database in detail … Tables
Where we need to work on more Research Areas 이동주 명재석 • Query Management • Parser • Query Expansion • Using dictionary, context • Query Mediation • Service Combining • Hot to combine different services? • When different services should be combined? • * Context Monitoring • Keeps monitoring on data & logs • Generate service rules • Push requests to matcher by catching context • Data Monitoring • Log Monitoring • Service Rule Generation • Matcher • Recieves requests • Decides which algorithms to be used Provides what the algorithms need • Recommendation • Music Recommendation • Other contents? • Question & Answering • Semantic Search • * Log Management • What do we need to log? • Data(Ontology) Modeling • User • Contents • Context • Ontology Manager • Lets us see heterogeneous data, logs, etc (physical) as one integrated ontology that includes user, context, contents,… (logical) • Methodology : how to map heterogeneous databases to an ontology? • Configurations • Inference Engine • Knowledge Expansion • Add semantics to raw data • Inference new semantics from existing context • * Mediator (Dealing with heterogeneity) • Data Mediator • Log Mediator • Database to Ontology Mapping • Synchronization • Index Manager • Crawler • Crawls data from a wide range of sources and saves to databases properly • Focused Crawling • Opinion Mining 강승석 이재원 이철기 이동주 양정연 Many of us