Submittion Number: [ Research Papers #2040 ]
Submittion Title: User-centric Unification of Web-scale Search and Reasoning from the Viewpoint of Granularity

Since in many cases, the authors in the field of AI do not write their papers whose titles include the exact words ``Artificial Intelligence'', they may mention more specific titles such as ``Agent'', ``Machine Learning'', etc. If more time is given, we may provide answers with a finer level of specificity according to the semantics and the hierarchical knowledge structure of ``Artificial Intelligence''.

We build an "Artificial Intelligence Ontology" by ourselves since this ontology contains all the possible branch information that has ever been proposed as sections or subsections in AI related conferences in the DBLP dataset and is relatively complete.

In most cases, the section names in the conference proceedings or journals are organized based on sub-branches of a field. Hence, we can extract the section and sub-section names to build the knowledge structure of a field. Because the SwetoDBLP dataset does not contain such kind of information, based on the XML version of the DBLP dataset (more specifically, the file dblp_bht.xml is used), we extract the branch information from ``Artificial Intelligence'' related conference proceedings and journals and build a three-level knowledge hierarchy.

In the knowledge hierarchy, Level~1 is ``Artificial Intelligence''. Levels~2 and 3 are extracted based on the tags <h2>*</h2> and <h3>*</h3> in the file dblp\_bht.xml. If the user allows more time, we can provide the number of authors in each sub-theory extracted between the tags <h2>*</h2> and an extended result list of authors in the field of AI. Further, if the user wants to get results finer than Level~2, then the subtheories which are between the tags <h3>*</h3> are extracted to build Level~3 in the knowledge hierarchy of AI. Thus, a new list of results according to sub-theories in Level~3 and a new extended result list of authors in the field of AI are provided.

In order to generate more relevant sub-theories for the specified field, we manually removed some not very relevant sources.Some supporting materials concerning the removed keywords can be found from [Here].