WO2011053112A1

WO2011053112A1 - System and method for visual query of semantic information

Info

Publication number: WO2011053112A1
Application number: PCT/MY2010/000218
Authority: WO
Inventors: Arun Anand Sadanandan; Kow Weng Onn; Mohammad Reza Beik Zadeh; Dickson Lukose
Original assignee: Mimos Berhad
Priority date: 2009-11-02
Filing date: 2010-10-22
Publication date: 2011-05-05

Abstract

The present invention relates generally to a system and method that enables a user to query information stored in knowledge bases by means of visual techniques which allows a user-friendly way of expressing queries especially when it involves complex queries with many parameters and conditions.

Description

SYSTEM AND METHOD FOR VISUAL QUERY OF SEMANTIC

INFORMATION

TECHNICAL FIELD OF THE INVENTION

BACKGROUND OF THE INVENTION

Retrieval of information is a vital aspect of the Semantic Web. It is no wonder that several user interface technologies have since been introduced and developed to assist users in accessing, retrieving and exploring the structured data stored in knowledge bases which is not only huge but complex as well. Retrieving information from any Resource Description Network (RDF) or Web Ontology Language (OWL) ontology requires a user to have in-depth knowledge of specialized query languages such as SPARQL or PROLOG because queries to the knowledge bases require one to follow the schema of tables or underlying data model. Query languages are usually rigid so one has to be familiar with the underlying data model or schema and layout of data across different tables as they require the users to precisely map their data needs to the underlying data model or schema. Furthermore, even in the case of a user being proficient in these languages, it is a challenging task to make complex and large queries due to the complexity of information stored in the knowledge bases and often there may be a mismatch of intended query and the database schema or failure to obtain an answer whereupon the user is forced to re-frame his entry based on the actual schema and contents of the knowledge base thus is time consuming and creates inefficiency.

There are times when queries are hard to express in the natural language form, especially when complex and lengthy information needs to be retrieved. Representing these kinds of queries in textual form could often lead to non-user friendly query structures, which can cause confusion even to human users not to mention to machines.

It would hence be extremely advantageous if the above shortcoming is alleviated by having a system where one does not need to learn and master the technical language specifications required to query any ontology and the answers retrieved from the knowledge bases can be converted into result graphs as well as readable text formats. This is disclosed in the present invention, which enables the user to construct graphs representing queries and the graph language is automatically converted into specific syntax structures to perform queries on ontologies.

SUMMARY OF THE INVENTION

Accordingly, it is the primary aim of the present invention to provide a system and method for visual query of semantic information, which allows a user-friendly way of expressing queries even when it involves complex queries with many parameters and conditions.

It is yet another object of the present invention to provide a system and method for visual query of semantic information which allows a user to interact with the concepts and properties present in the ontology thereby contributing to the efficient design of accurate queries. Yet another object of the present invention is to provide a system and method for visual query of semantic information wherein the process of creating queries is simplified for easy to manipulation.

Yet a further object of the present invention is to provide a system and method for visual query of semantic information, which requires no in- depth knowledge in the query languages involved or inner workings of the database thereby allowing the non-technical community and novices to effectively build queries and use the system effectively.

It is a further object of the present invention to provide a system and method for visual query of semantic information, which allows the user to access database without having to learn the complicated query languages and graphical representations of queries.

Yet a further object of the present invention is to provide a system and method for visual query of semantic information, which allows syntax structure validity, improved efficiency and understanding as well as reduced training requirements.

Yet another object of the present invention is to provide a system and method for visual query of semantic information wherein the answers retrieved based on queries by the user is in the readable form including textual representation and result graphs.

Other and further objects of the invention will become apparent with an understanding of the following detailed description of the invention or upon employment of the invention in practice.

According to a preferred embodiment of the present invention there is provided,

A visual semantic query system comprising; at least a knowledge base (2); at least a query formulation module (6); at least a result visualiser module (12) characterised in that said query formulation module (6) is linked to at least a visual query language interpreter module (8) and the said knowledge base (2); further characterised in that said visual query language interpreter module (8) is linked to at least a query processor module (10) and said knowledge base (2).

In another aspect, the present invention provides,

A method of performing visual query on semantic information comprising steps of, loading the ontology; constructing query formulation by interacting with the concepts and properties of the ontology to create a query graph; automatically mapping of the query graph onto syntactically valid internal structure to generate visual query language; executing the automatically generated visual query language and retrieving the results; processing the results to generate readable answer forms; creating representation of the answers in textual or visual form. BRIEF DESCRIPTION OF THE DRAWINGS

Other aspect of the present invention and their advantages will be discerned after studying the Detailed Description in conjunction with the accompanying drawings in which:

FIG. 1 is a block diagram showing the architecture of the visual semantic query system in its preferred embodiment.

FIG. 2 is a flow chart detailing the graph query construction process.

FIG. 3 is a flow chart illustrating an operational sequence for the visual semantic query process carried out in the visual semantic query system illustrated in FIG. 1.

DETAILED DESCRIPTION OF THE DRAWINGS

In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be understood by those of ordinary skill in the art that the invention may be practised without these specific details. In other instances, well known methods, procedures and/ or components have not been described in detail so as not to obscure the invention.

The invention will be more clearly understood from the following description of the embodiments thereof, given by way of example only with reference to the accompanying drawings, which are not drawn to scale.

Referring to FIG. 1, there is shown a block diagram of the architecture of the visual semantic query system in its preferred embodiment comprising at least a knowledge base (2), at least an ontology loader module (4), at least a query formulation module preferably a query visualiser module (6), at least a visual query language interpreter module (8), at least a query processor module (10) and at least a result visualiser module (12). Visual Query Systems (VQSs) are techniques supporting query formulation using visual means that is it is a system for information retrieval using a visual representation to depict the domain of interest and expressed related requests. This system allows users to express queries visually and provide tools to allow users to interact with the system. The current invention has a graphical User Interface (GUI) where the input comes in the form of user interaction with the graphical elements. The output of the system is in at least two modalities namely text form and visual form (result graphs). The target users for these systems are typically non-technical people and novices, who are not concerned about the inner structure of the database. Having said that, the VQSs could also benefit expert users, especially when it comes to dealing with complex queries and data structures.

The ontology loader module (4) is invoked when the user initially selects ontology to be loaded. In this module a list of concepts and properties in the ontology is loaded into the interface and is derived from establishing connections with the knowledge base (2) as indicated by the arrow "A".

The next phase after ontology loading is the query formulation phase. In this phase the query visualizer module (6) is invoked where query formulation in the form of graphs using nodes and arcs are constructed. In this module the concepts and properties of the selected ontology are populated on the Graphical User Interface (GUI) for further processing. Concepts and properties are also additionally derived from establishing connections and interactions with the knowledge base (2) as indicated by the arrow "B" to populate the GUI for further processing. As the query graph is being constructed more concepts and properties may be retrieved from the knowledge base (2). Since the system supports data in RDF format, the query graphs are based on the subject-predicate-object notation of RDF, where the subjects and objects are represented by nodes and the predicates are represented as edges between the subjects and objects. The user has the option of starting the query graph construction process using the Visual Query language through two methods which will described hereinbelow and as illustrated in FIG. 2 below.

Once the user has constructed the query graph, a visual query language interpreter module (8) is employed. The main function of this module is to do an automatic mapping of the query graph into a syntax specific textual representation language such as SPARQL or Prolog. In other words this module analyzes the query graph by converting the concepts and properties of the query graph into equivalent statements in SPARQL or Prolog. This is achieved by first converting the graph into the subject-predicate-object form and then attaching the syntax specific elements automatically. Additionally certain solution modifiers (if applicable) are also added to the query format.

Subsequently a query processor module (10) is invoked. This module is concerned with establishing connections with the knowledge base (2) as indicated by the arrow "C" and executing the query statements created by means of the previous modules. This is a generic module, which can be replaced with other knowledge base processor engines.

Once the query processor module (10) receives the answers from the knowledge base (2), the next step is to convert the syntax specific results into readable answer forms and display the answers to the user in graphical form or other readable textual form. These functionalities are performed in the result visualiser module (12). This module can be extended to render the results for the user in several other modalities such as charts, presentations, external applications and etcetera.

Referring now to FIG. 2, there is shown a flow chart detailing the graph query construction process comprising several graph query construction steps namely a node creation step indicated by the reference numeral (14), node selection step indicated by the reference numeral (16), a node ascertainment step indicated by the reference numeral (18), a property display step indicated by the reference numeral (20), a property addition step indicated by the reference numeral (22), a query element addition step indicated by the reference numeral (23). The user has the option of starting the query graph construction process using the Visual Query language through two methods wherein the first method is to create or construct a query graph using a known concept whilst the second method is to create or construct a query graph from an unknown concept.

In the first method of creating or constructing a query graph using a known concept and as illustrated in FIG. 2, the user starts by creating or adding a node or arc to the query workspace as shown in the node creation step (14). In the node selection step (14), the user then selects any node. When a node has been selected, the next step, the node ascertainment step (18), will be invoked to ascertain the type of node that has been selected. This node ascertainment step (18) consists of a few sub- node ascertainment steps to assist in ascertaining the type of node that has been selected. In the first sub-node ascertainment step (18A), it will be ascertained whether the selected node is a root node. If the selected node is a root node then it will be advanced to the second sub-node ascertainment step (18B) to further determine whether the node is a known concept or not. If it is an unknown concept (blank node), all properties are shown in the list of properties as shown in the property display step (20). The required property is then added to the selected node in the property addition step (22) to construct the query graph. The next step is the query element addition step (23) where the user is allowed to decide whether to add additional query elements in the construction of the query graph or not. Following this, if the user desires to continue adding more properties and concepts to extend the query graph to make complex queries, the whole process will be repeated as indicated by the arrow "D". On the other hand, if it is a known concept then all properties of that concept are shown as a list of properties for the user to browse or search through as shown in the fourth sub-node ascertainment step (18D). In this fourth sub-node ascertainment step (18D) all the properties, which are attached to that particular concept, are filtered in a properties panel and this filtering is the result of the execution of a SPARQL query snippet generated from the selected concept. The required property is then added to the selected node in the property addition step (22) to construct the query graph and the steps that follow are similar to the steps described earlier (that is if found to be an unknown concept).

If the selected node is firstly found not to be a root node in the first sub-node ascertainment step (18A) then all the concepts in the range of property attached to the concept is filtered in the list of concepts.

In the second method, the user starts by creating or constructing a query graph from an unknown concept indicated by a blank node. This method can be used if the user has information only about the properties of the query is known. Referring to FIG. 2 again, the user starts by creating or adding a node to the query workspace as shown in the node creation step (14). Then the user selects any node; in this case a blank node [as the concept is unknown to the user but only information about the properties of the query is known]. When such a node is selected, the next step, the node ascertainment step (18), will be invoked to ascertain the type of node that has been selected. This process is similar to the one carried out in the first method. In the first sub-node ascertainment step (18A), it will be ascertained whether the selected node is a root node. If the selected node is a not a root node then it will be advanced to the third sub-node ascertainment step (18C) where all the concepts in the range of the property attached to the concept is filtered in the list of concepts.

Although the concept is unknown, the property attached to it is known, so the list of concepts (which fall in the range of property) is populated on the query graph based on the property attached. Once that is done the node will be advanced to the second sub-node ascertainment step (18B) to ascertain whether the node is a known concept or not. If the selected node is not a known concept, then all the properties of the selected concept are filtered in the list of properties in the fourth sub-node ascertainment step (18D). Therefore selecting a blank node will enable filtering of a range of concepts that the property can be assigned. The required property is then added to the selected node in the property addition step (22) to construct a query graph. After this the user can continue extending the graph to make complex queries by repeating the whole process as indicated by the arrow "D".

Alternatively, the user can load a pre-stored graph or save the constructed graph for later use.

Referring now to FIG. 3, there is shown a flow chart illustrating an operational sequence for the visual semantic query process carried out in the visual semantic query system illustrated in FIG. 1 comprising six major steps namely an ontology loading step indicated by the reference numeral (24), a graph query construction step indicated by the reference numeral (26), an automatic graph query mapping step indicated by the reference numeral (28), a query executing step indicated by the reference numeral (30), a query processing step indicated by the reference numeral (32) and a visual representation step indicated by the reference numeral (34). Once the user logs into the system, the first step is to load the ontology, which is derived from the knowledge base (2), which is carried out in the ontology loading step (24). Then the user constructs query formulation preferably a query graph by selecting and adding required concepts and properties from the loaded ontology data in the graph query construction step (26). As the query graph is being constructed in the graph query construction step (26) more concepts and properties may be retrieved from the knowledge base (2) directly as indicated by the arrow "B" in FIG. 1. The query graph is subsequently interpreted by automatically mapping the query graph with syntactically valid internal structure to create an internal representation language (or also referred to as query statement or visual query language) such as SPARQL or Prolog that is required for retrieving answers from the knowledge base (2) in the automatic graph query mapping step (28). Once the graph is constructed and completely interpreted the automatically generated query statement is then executed to retrieve results from the knowledge base (2) in a query- executing step (30). The retrieved answers or results which are in a specific format are then processed and converted into readable forms in at least two modalities namely textual representation and visual representation such as result graphs in a query processing step (32) which is carried out in the query processor module (10). Additionally the system also supports utility functions to load or save graphs and to perform regular graph operations such as adding nodes, deleting nodes and so on.

While the preferred embodiment of the present invention and its advantages has been disclosed in the above Detailed Description, the invention is not limited thereto but only by the spirit and scope of the appended claim.

Claims

WHAT IS CLAIMED IS:

1. A visual semantic query system comprising; at least a knowledge base (2); at least a query formulation module (6); at least a results visualiser module (12) characterised in that said query formulation module (6) is linked to at least a visual query language interpreter module (8) and the said knowledge base (2); further characterised in that said visual query language interpreter module (8) is linked to at least a query processor module (10) and the said knowledge base (2).

2. A visual semantic query system as in Claim 1 wherein the query formulation module (6) facilitates the query construction by means of graphs using nodes or arcs based on the interaction of the user with the concepts and properties of the loaded ontology that is inputted in the ontology loader module (4).

3. A visual semantic query system as in Claim 1 wherein the query formulation module (6) facilitates the formulation and construction of query graphs based on the interaction of the user with the concept and properties of the knowledge base (2).

4. A visual semantic query system as in Claim 1, which enables the users to query information stored in knowledge bases in RDF format

5. A visual semantic query system as in Claim 1 wherein the visual query language interpreter module (8) carries out automatic mapping of the query graph into a syntax specific textual representation language.

6. A visual query language interpreter module (8) as in Claim 1 or 5 wherein the textual representation language employed is SPARQL or Prolog.

7. A visual semantic query system as in Claim 1 wherein the system is further characterised by the provision of a query processor (10) to automatically convert the syntax specific results or answers pursuant to execution of query statements into readable forms.

8. A visual semantic query system as in Claim 1 wherein the system is further characterised by the provision of a visualiser module (12) that is capable of displaying readable answer forms to the user.

9. A visual semantic query system as in Claim 8 wherein the visualiser module (12) is capable of creating textual or visual representations of the results or answers retrieved.

10. A visual semantic query system as in Claim 9 wherein the visualiser module (12) is capable of creating visual representations in the form of graphical representation, charts or presentations.

11. A method of performing visual query on semantic information comprising steps of, loading the ontology; constructing query formulation by interacting with the concepts and properties of the ontology to create a query graph; automatically mapping of the query graph onto syntactically valid internal structure to generate visual query language; executing the automatically generated visual query language and retrieving the results; processing the results to generate readable answer forms; creating representation of the answers in textual or visual forms.

A graph query construction process where the graph is constructed from a known concept comprising steps of, creating a node to the query workspace; selecting a node; ascertaining the type of node selected to determine whether it is a root node; ascertaining the type of node selected to determine whether it is a known concept if it is ascertained to be a root node; displaying all the properties, which are attached to a particular concept if has been ascertained to be a known concept; adding a property to the selected node; deciding whether to add additional query elements to the query graph or not.

A graph query construction process where the graph is constructed from an known concept comprising steps of, creating a node to the query workspace; selecting a node; ascertaining the type of node selected to determine whether it is a root node; filtering all the concepts in the range of the properties of the query sought for into a list of concepts if it is ascertained not to be a root node; ascertaining whether it is a known concept; displaying all the properties, which are attached to a particular concept if has been ascertained to be a known concept; adding a property to the selected node; deciding whether to add additional query elements to the query graph or not.

14. A graph query construction process where the graph is constructed from an unknown concept comprising steps of, creating a node to the query workspace; selecting a node; ascertaining the type of node selected to determine whether it is a root node; ascertaining the type of node selected to determine whether it is a known concept if it is ascertained to be a root node; ascertaining whether it is a known concept; filtering all the properties of the selected concept in the list of properties if it is ascertained not to be a known concept; adding a property to the selected node; deciding whether to add additional query elements to the query graph or not.

A graph query construction process where the graph is constructed from an unknown concept comprising steps of, creating a node to the query workspace; selecting a node; ascertaining the type of node selected to determine whether it is a root node; filtering all the concepts in the range of the properties of the query sought for into a list of concepts if it is ascertained not to be a root node; ascertaining whether it is a known concept; filtering all the properties of the selected concept in the list of properties if it is ascertained not to be a known concept; adding a property to the selected node; deciding whether to add additional query elements to the query graph or not.