Systems
ARISE has developed some entity information search service for the restaurant domain. The service is published as both iOS mobile app and general Web service, as shown in the Figure 1 and Figure 2.
Figure 1. iOS version restaurant search service.
Figure 2. Web version restaurant search service.
ARISE is now building some entity information service for the people domain. Our core technology comprises data mining and Web search for extraction, organization and retrieval of entity-oriented information from the Web. Figure 3 demonstrates the architecture of our ARISE system. Consider an example in virtual meeting, where a user wants to browse the information about the meeting participants. At step 1, the user can initiate an entity query through our system API. Then at step 2, the query is parsed by our API and used to quickly search our entity database. If the person matches any record, then the information will be returned at step 3 based on some well-designed entity ranking function.
Figure 3. System architecture of ARISE people entity information service platform
If the person does not match any record in our database (i.e. step 4), then we will initiate an extraction-integration process based on our core technology. Generally, the entity query is used for Web search, in order to harvest some Web pages about the interested person entity (i.e. step 5). Then, the crawled Web pages are carefully parsed and fed into our entity extraction component to extract the relevant entity information (i.e. step 6). After the extraction, we can further integrate the entity information (i.e. step 7). After this integration is done, the well-organized information is saved into our database and indexing engine (i.e. step 8). Finally, the database and indexing engine can further serve the API.
In addition to the general entity query search, our system can also serve other applications which require real-world entity information. For example, a location-based advertising application can use our restaurant entity information to help decide what advertisements to show in different locations based on some carefully designed auction mechanism.