WO2004090808A2

WO2004090808A2 - Computer-implemented system and method for progressively transmitting knowledge and computer program product related thereto

Info

Publication number: WO2004090808A2
Application number: PCT/EP2004/050487
Authority: WO
Inventors: Maria Athelogou; Ursula Benz
Original assignee: Definiens Ag
Priority date: 2003-04-07
Filing date: 2004-04-07
Publication date: 2004-10-21
Also published as: US20060026111A1; WO2004090808A8; EP1611547A2; US8909692B2

Abstract

A computer-implemented system for progressively transmitting of knowledge between system nodes (4, 5, 7) of a network structure comprises a plurality of system nodes (4, 5, 7) and intelligent interfaces by which respective system nodes (4, 5, 7) are coupled with each other for performing a communication. The intelligent interfaces transmit object features of cognition structure objects comprising knowledge, information and data depending on a respective question of a respective one system nodes (4, 5, 7) progressively more faithful to detail from another of the respective system nodes (4, 5, 7) to the one of the respective system nodes (4, 5, 7). Furthermore, there are disclosed a corresponding method and a computer program product relating to the system and method.

Description

TITLE OF THE INVENTION

COMPUTER-IMPLEMENTED SYSTEM AND METHOD FOR PROGRESSIVELY TRANSMITTING KNOWLEDGE AND COMPUTER

PROGRAM PRODUCT RELATED THERETO

TECHNICAL FIELD

The present invention relates to an computer-implemented system and method for progressively transmitting knowledge and a computer program product related thereto.

BACKGROUND ART

In the art there are known computer systems in which generating, storing and/or processing of data does not take place at the same location. Therefore, an amount of data to be transmitted forms a bottle-neck despite increasing resources and can lead to a significant restriction of efficiency of the whole system.

To reduce the amount of data to be transmitted there are known various data compression methods. It can generally be distinguished between methods using lossy source coding and methods using loss-free source coding. Source coding is conversion of a data word in a corresponding data word using of a redundancy decreasing code. In loss-free source coding transmitted data can be fully restored by a suitable decompression, whereas this is not the case in lossy source coding.

Currently used adaptive data compression methods are for example JPEG and MPEG, wherein these are applied to all data of a data set. These adaptive data compression methods fit on a mostly fluctuating statistic of the data. Therefore, these methods are often called "context sensitive". Either the transmitter or the receiver determines a medium or minimal quality of reconstruction for all data or a data rate before transmission, respectively . All data are then transmitted in correspondence with the determined quality or data rate, respectively.

Furthermore, there are known so-called progressive data compression methods that usually are based on so-called wavelets. These progressive data compression methods allow for a piece by piece reconstruction of data, wherein the receiver can determine which transmitted amount of data is sufficient. During transmission it can be determined when transmission of data is canceled. Therefore, an amount of data to be transmitted can be decreased. In general, all date are transmitted until transmission of data is canceled.

Independent of data compression data mining methods, i.e. methods that relate to a directed or subject-related search of data, allow that data sets of interest are searched for in a database. Here, the term "of interest" is limited to identification of data sets having similar features. Detection of data sets of interest normally takes place by monitored training using examples.

However, it is common to all aforementioned methods that on the one hand finally merely data are transmitted and on the other hand compression of data is limited to such a degree that the receiver can meaningful process and/or use transmitted data. The respective used transmission channel is consequently not used effectively, thereby leading to a bottle-neck on the transmission channel.

DISCLOSURE OF THE INVENTION

It is accordingly the object of the present invention to provide a computer- implemented system and method by means of which a more effective use of a respective transmission channel is possible.

With respect to the system this object is solved by the measures indicated in claim 1 , with respect to the method this object is solved by the measures indicated in claim 8 and with respect to the computer program product this object is solved by the measures indicated in claims 15 and 16.

Further advantageous modifications of the present invention are subject matter of the dependent claims.

In more detail, according to a first aspect of the present invention there is provided a computer-implemented system for progressively transmitting of knowledge between system nodes of a network structure, comprising a plurality of system nodes, and intelligent interfaces by means of which respective system nodes are coupled with each other for performing a communication therebetween, wherein the intelligent interfaces transmit object features of cognition structure objects comprising knowledge, information and data depending on a respective question of a respective one of the system nodes progressively more faithful to detail^' from another of the respective system nodes to the one of the respective system nodes.

In more detail, according to the first aspect of the invention roughly focussed knowledge, information and/or data is transmitted with highest priority. Such roughly focussed knowledge, information and/or data occupies much less transmission capacity of a transmission channel than data itself.

Only when it is determined that further detailed knowledge, further detailed information and/or further detailed data is necessary, these are transmitted with a priority corresponding to the aforementioned order. In this manner there is provided a progressive transmission of knowledge in which the transmission channel is more effectively used as compared with conventional data compression methods.

According to a second aspect of the present invention there is provided a computer-implemented method for progressively transmitting knowledge between system nodes of a network structure, the method comprising the step of progressively more faithful to detail transmitting object features of cognition structure objects comprising knowledge, information and data depending on a respective question of one of the respective system node from another of the respective system node to the one of the respective system nodes.

BRIEF DESCRIPTION OF DRAWINGS

These and other aspects of the present invention will be apparent from and elucidated with reference to the embodiment described hereinafter and considered in connection with the accompanying drawings, in which:

Figs. 1A and 1 B are a schematic illustration of a knowledge base which can be used in an embodiment of the present invention in two different kinds of representation; Fig. 2 is an illustration of a relation between data, information and knowledge according to the embodiment of the present invention;

Fig. 3 is a diagram of membership functions used in the embodiment of the present invention;

Fig. 4 is an illustration of a structure of a system according to the embodiment of the present invention;

Figs. 5 to 11 are schematic illustrations of different processing operations according to the embodiment of the present invention;

Figs. 12 to 14 are flow charts concerning the different processing operations shown in Figs. 5 to 11 ; and

Fig. 15 is a overall structure of a generalized system according to the embodiment of the present invention.

BEST MODE FOR CARRYING OUT THE INVENTION

In an embodiment of the present invention described hereinafter a self- organizing, semantic, self-similar cognition network of objects or triple-SN, respectively can be used as a knowledge base. Such a triple-SN is for example described under the designation "fractal-hierarchical network" in WO00/20964, WOOO/63788, WOOO/00497, WO01/45033 and WO02/05198, respectively, wherein methods described in these documents can be furthermore used in a processing operation according to the embodiment of the present invention. These documents are incorporated herein by reference.

Figs. 1A and 1 B are a schematic illustration of a knowledge base which can be used in an embodiment of the present invention. Figs. 1A and 1B merely refer to two different kinds of representation and the section shown in Fig. 1A is not identical to the section shown in Fig. 1 B.

As has been already mentioned above, the knowledge base is constituted in form of a self-organizing, semantic, self-similar cognition network of objects or triple-SN, respectively. Such a triple-SN consists of semantic units 1 which are connected with each other via linking units 3 as shown by the connections between the semantic units 1 in Figs. 1A and 1B. By multiple and repeated merging of respective semantic units 1 of a lower hierarchical level to a super- ordinated semantic unit 1 of a higher hierarchical level there is generated a hierarchical network of world knowledge. In case of this triple-SN the world knowledge is in more detail fractal-hierarchical, because similar processing operations are applied to all hierarchical levels. The triple SN provides for a hierarchical structure of an arbitrary dimensionality, i.e. an n-dimensional object structure, wherein a respective hierarchical level of the hierarchical structure corresponds to a certain resolution of the world knowledge, which becomes more detailed in the hierarchical structure from the top to the bottom.

An essential component of the triple-SN is a particular form of the semantic unit 1 , i.e. a so-called Janus-unit or processing unit 2, respectively. By a semantic Janus-unit 2 which is appended to one or more semantic units 2 local operations can be performed in the triple-SN. These local operations comprise inter alia generating of new semantic units 1 , merging of already existing semantic units 1 to a single semantic unit 1 which is newly generated where required, modifying or deleting of already existing semantic units 1 and comparing of already existing semantic units 1. In this manner of using Janus-units 2 existing n the triple-SN it is possible to modify the world knowledge in all possible manners.

The linking units 3 mentioned above can also be a particular form of the semantic units 1. In this manner, it is possible to perform the aforementioned arbitrary operations between arbitrary types of semantic units 1 which also comprise operations on Janus-units 2 and/or linking units 3 and therefore provide a triple-SN which can be arbitrarily modified. Therefore, the knowledge base existing in the triple-SN and consequently all in all the knowledge base can be arbitrarily modified.

In general, it can be considered that the triple-SN comprises a network of structure objects and a network of class objects. Structure objects are derived from respective data, information and/or knowledge. Class objects represent classes to which structure objects can be classified. Structure objects classified to respective class objects therefore represent respective instances of the respective class objects. The linking units 3 represent relations between respective of the above structure objects and/or class objects. Linking units 3 between respective structure objects on a same hierarchical level of the triple-SN represent neighborhood relations between the respective structure objects. Linking objects 3 between respective structure objects on different hierarchical levels of the triple- SN represent sub-ordinated and super-ordinated relations between the respective structure objects. Linking objects 3 between respective class objects on a same hierarchical level of the triple-SN represent neighborhood relations between the respective class object. Linking objects 3 between respective class objects on different hierarchical levels of the triple-SN represent sub-ordinated and super- ordinated relations between the respective class objects Finally, linking objects 3 between respective structure objects and class objects represent classifying relations between the respective structure objects and class objects.

Although the embodiment of the present invention preferably uses the triple- SN as a knowledge base there is the possibility to use another hierarchical structure in which relations are present between respective objects.

Before the structural configuration of the embodiment of the present invention is described in detail, there is described in which manner the relation between data, information and knowledge is defined and in which manner information can be derived from data and knowledge can be derived from information.

Fig. 2 is an illustration of a relation between data, information and knowledge according to the embodiment of the present invention.

It should be noted that objects containing at least one of knowledge, information or data are hereinafter called "cognition structure objects".

According to the embodiment of the present invention feature profiles comprising features and membership functions are used to derive Information from data as shown in Fig. 2. Furthermore, rules are used to derive knowledge from the information derived from the data. This will become obvious in more detail from the following description of the structure of a knowledge base used in the embodiment of the present invention.

As already described above, the knowledge base of the embodiment of the present invention is constructed as a triple-SN. Each semantic unit within the triple-SN is described by the relations to its sub-ordinate objects, to its neighborhood and to its super-ordiπate objects, by its features and by its feature profiles. On the lowest hierarchical level of the triple-SN the data are stored which further comprise the spatial position as a feature. The lowest hierarchical level does not comprise sub-ordinate objects in the embodiment of the present invention. However, there could be a hierarchical level k = -1 with a separation if a end-member analysis is performed. The highest hierarchical level does not comprise super-ordinate objects.

In the hierarchical structure of the triple-SN each hierarchical level of the triple-SN is defined by a number k of the hierarchical level and a feature vector

F(k) which comprises quantities characterizing a totality of all objects on the hierarchical level k such as medium object size, number of objects, minimal and maximal object size, strength of segmentation of objects and so on.

An example of such feature vector F(k) is mentioned below in equation (1 ).

Furthermore, for each hierarchical level k of the triple-SN there is defined a feature matrix F'(k) which is a matrix of degrees μp/| (B) of membership to N base classes B and the degrees μp(k)(^B) o^f membership are derived from membership functions for this hierarchical level k. One possible example of such a feature matrix F'(k) is shown below in equation (2).

The membership functions μp(k)(^B) to the N base classes define the feature profile of a feature of the hierarchical level k. There is also the possibility to store merely the membership functions instead of the degrees μp | (B) in the feature matrix F'(k). The degrees μp(k)(B) of membership are then calculated if necessary. This leads to a higher flexibility because new membership functions can be transmitted and stored to further adapt an evaluation to a specific application.

Fig. 3 is a diagram of membership functions used in the embodiment of the present invention.

In more detail, Fig. 3 shows a feature profile of a feature x. Th abscissa shows the co-domain of feature x and the ordinate shows the degree of membership of feature x to co-domains B1 to B3 which constitute classes in the embodiment of the present invention.

Therefore, according to this embodiment of the present invention feature profiles comprising relevant features and relevant membership functions are defined for data. These feature profiles again define a multi-dimensional space in which the original data are transformed to minimize an amount of transmission. This transformation generates information from the data.

Furthermore, there are defined rules by means of which information derived from data with the aid of the feature profiles can be assigned to classes. These rules represent a further transformation prescription which transforms information to knowledge which describes to which degree respective data can be assigned to a class which is searched for.

As a result, there is provided a triple-SN in which data, information and knowledge is stored in semantic units which are also called "cognition structure objects" in the embodiment of the present invention.

Fig. 4 is an illustration of a structure of a system according to the embodiment of the present invention.

As shown in Fig. 4 the system of the embodiment of the present invention comprises a client 4, a sensor 5, a database 6 and a sensor 7 and is used to view a scene 8 comprising different objects 9. It should be noted that the present invention is not only be applicable to visual scenes. Rather, the present invention can be applied to any kind of cognition structure objects such as text data, spatial data, time data and so on simultaneously, successively, dynamically, locally adapted and so on and also be used to generate any kind of cognition structure objects simultaneously, successively, dynamically, locally adapted an so on. There is also the possibility that cognition structure objects are transformed from one kind to another kind. This means that different kinds of cognition structure objects can be arbitrarily processed. Thus, the present invention provides a multi-modal system and method. Furthermore, the database 6 is constituted by the knowledge base described above. Furthermore, the client 4, the server 5 and the sensor 7 can generally be called "system nodes" because the manner in which transmission of object features of cognition structure objects is identical for all system nodes.

In addition to the database 6 at the side of the server 5 there is also the possibility that cognition structure objects can be locally stored at the side of the client 4 or at the side of the sensor. For example, the sensor could detect and create cognition structure objects based on the objects 9 within the scene 8 before a question as to specific cognition structure objects is made for example from the client 4 to the sensor 7. In this case the sensor 7 could firstly use the cognition structure objects stored therein and thereafter use cognition structure objects newly detected and created based on the objects 9 within the scene 8 with respect to the question from the client 4. Further such additional modifications of the present invention are obvious for the man skilled in the art.

As is obvious from the arrows in Fig. 4 there are different possibilities to communicate between the client 4, the server 5, the sensor 7 and the database 6 as will be described below. These will be described below in further details.

In the following illustration the database 6 is not shown but it is assumed that the database 6 is also connected to the server as shown in Fig. 4.

Fig. 5 is a schematic illustration of a first processing operation according to the embodiment of the present invention. In more detail, Fig. 5 shows an interaction between the client 4 and the server 5. This interaction is described in more detail with reference to Fig. 12.

As is shown in Fig. 12 in a step S100 there is checked whether or not a question regarding data, information and/or knowledge stored in the database 6 is transmitted from the client 4 to the server 5. If no question is transmitted the procedure returns to the step S100. If a question is transmitted in the step S100 the procedure proceeds to a step S110 in which it is checked whether or not a relevant cognition structure comprising data, knowledge and/or information is extractable from the database 6. This is done at a high hierarchical level of the triple-SN constituting the knowledge base and consequently the database 6 the cognition structure objects are present in a very low resolution. It should be noted that the cognition structure objects are arranged within the triple-SN progressively more faithful to detail from a higher hierarchical level of the triple-SN to a lower hierarchical level of the triple-SN.

If it is determined in a step S110 that no relevant cognition structure is extractable the procedure proceeds to the step S120 in which it is determined whether or not a relevant cognition structure is creatable. If it is determined in the step S120 that no relevant cognition structure is creatable the procedure ends. Otherwise, if it is determined in the step S120 that a relevant cognition structure is creatable the procedure proceeds to a step S130 in which the relevant cognition structure is created. Thereafter, the procedure proceeds to a step S140 in which the relevant cognition structure is outputted from the database 6 to the client 4 via the server 5. After the step S140 the procedure proceeds to a step S170 in which it is determined whether or not the question has been fully answered. If it is determined in the step S170 that the question has been fully answered the procedure ends. Otherwise, if it is determined in the step S170 that the question has not been fully answered the procedure returns to the step S110 and is than again performed at a lower hierarchical level of the triple-SN to obtain a more detailed cognition structure of the triple-SN to fully answer the question.

If it is determined in the first determination of the step S110 that a relevant cognition structure is extractable the procedure proceeds to the step S150 in which the relevant cognition structure is extracted form the database 6. Thereafter, the procedure proceeds to a step S160 in which the relevant cognition structure is outputted from the database 6 to the client 4 via the server 5. In the following step S170 it is checked whether or not the question has been fully answered. If it is determined in the step S170 that the question has been fully answered the procedure ends. Otherwise, if it is determined in the step S170 that the question has not been fully answered, the procedure returns to the step S110 and is than again performed at a lower hierarchical level of the triple-SN to obtain a more detailed cognition structure of the triple-SN to fully answer the question.

As is obvious from Fig. 12, the procedure is iteratively performed from higher hierarchical levels to lower levels of the triple-SN until the question has been fully 5 answered, no relevant cognition structure is extractable or creatable or the procedure is interrupted or terminated for example by a user and so on.

By means of the aforementioned procedure it is possible to transmit object features of cognition structure objects existing in the database 6 from the same to 0 the client 4 via the server 5. Furthermore, new cognition structure objects can be created and outputted to the client 4. In addition there is also the possibility to store such new cognition structure objects at the side of the database 6 and to use such new cognition structure objects in further processing operations. Furthermore, all processing operations described above with respect to the triple- 5 SN can be performed at the time of performing the aforementioned procedure to modify cognition structure objects stored in the database. In addition, the object features of the extracted and/or created cognition structure objects can be additionally stored at the side of the client 4 to be available as local knowledge, information and/or data at a later processing operation. The same holds true in a o similar manner for all other procedures for transmission object features between system nodes such as client 4, server 5 and sensor 7.

Fig. 6 is a schematic illustration of a second processing operation according to the embodiment of the present invention. In more detail, Fig. 6 shows an 5 interaction between the client 4 and the sensor 7. As is shown in Fig. 13 in a step S200 there is checked whether or not a question regarding data, information and/or knowledge about the objects 9 within the scene 8 detected by the sensor 7 or regarding locally stored data, information and or knowledge at the side of the sensor 7 is transmitted from the client 4 to the sensor 7. If no question is 0 transmitted the procedure returns to the step S200. If a question is transmitted in the step S200 the procedure proceeds to a step S210 in which it is checked whether or not a relevant cognition structure comprising data, knowledge and/or information is extractable from the sensor 6 and/or the scene 8. This is done at a high hierarchical level. 5

If it is determined in the step S210 that no relevant cognition structure is extractable the procedure proceeds to the step S220 in which it is determined whether or not a relevant cognition structure is creatable. If it is determined in the step S220 that no relevant cognition structure is creatable the procedure ends.

Otherwise, if it is determined in the step S220 that a relevant cognition structure is creatable the procedure proceeds to a step S230 in which the relevant cognition structure is created.

Thereafter, the procedure proceeds to a step S240 in which the relevant cognition structure is outputted from the sensor 7 to the client 4. Thereafter, the procedure proceeds to a step S270 in which it is determined whether or not the question has been fully answered. If it is determined in the step S270 that the question has been fully answered the procedure ends. Otherwise, if it is determined in the step S270 that the question has not been fully answered the procedure returns to the step S210 and the same is again performed at a lower hierarchical level to obtain a more detailed cognition structure to fully answer the question.

If it is determined in the first determination of the step S210 that a relevant cognition structure is extractable the procedure proceeds to the step S250 in which a relevant cognition structure is extracted. Thereafter, the procedure proceeds to a step S260 in which the relevant cognition structure is outputted from the sensor 7 to the client 4. Thereafter, the procedure proceeds to the step S270 in which it is checked whether or not the question has been fully answered. If the question has been fully answered the procedure ends. If the question has not been fully answered the procedure returns to the step S210 and is than again performed at a lower hierarchical level of the triple-SN to obtain a more detailed cognition structure of the triple-SN to fully answer the question.

As is obvious from Fig. 13, the procedure is iteratively performed from higher hierarchical levels to lower levels until the question has been fully answered, no relevant cognition structure is extractable or creatable or the procedure is interrupted or terminated for example by a user and so on.

Fig. 7 is a schematic illustration of a third processing operation according to the embodiment of the present invention. In more detail, Fig. 7 shows an interaction between the sensor 7 and the server 5. This interaction is described in more detail with reference to Fig. 14. As shown in Fig. 14 in a step S300 there is checked whether or not a question regarding data, information and/or knowledge in the scene 8 at the side of the sensor and/or locally stored data, information and/or knowledge at the side of the sensor 7 is transmitted from the server 5 to the sensor 7. If no question is transmitted the procedure returns to the step S300. If a question is transmitted in the step S300 the procedure proceeds to a step S310 in which it is checked whether or not a relevant cognition structure comprising data, knowledge and/information is extractable from the sensor 7 or the scene 8. This is done at a high hierarchical level.

If it is determined in the step S310 that no relevant cognition structure is extractable the procedure proceeds to the step S120 in which it is determined whether or not a relevant cognition structure is creatable. If it is determined in the step S120 that no relevant cognition structure is creatable the procedure proceeds to the step S330 in which the relevant cognition structure is created.

Thereafter, the procedure proceeds to the step S340 in which the relevant cognition structure is outputted from the sensor 7 to the server 5. Thereafter, the procedure proceeds to a step S370 in which it is checked whether or not the question has been fully answered. If it is determined in the step S370 that the question has been fully answered the procedure ends. Otherwise, if it is determined in the step S370 that the question has not been fully answered the procedure returns to the step S310 and is than again performed at a lower hierarchical level to obtain a more detailed cognition structure to fully answer the question. If it is determined in the first determination of the step S110 that a relevant cognition structure is extractable the procedure proceeds to a step S150 in which the relevant cognition structure is extracted from the data base 6. Thereafter, the procedure proceeds to a step S360 in which the relevant cognition structure is outputted from the sensor 7 to the server 5. In the following step S370 it is checked whether or not the question has been fully answered. If the question has been fully answered the procedure ends. If the question has not been fully answered the procedure returns to the step S 10 and is than again performed at a lower hierarchical level of the triple-SN to obtain a more detailed cognition structure of the triple-SN to fully answer the question.

As is obvious from the procedure shown in Fig. 14 the procedure is iteratively performed from higher hierarchical levels to lower hierarchical levels of the triple- SN. The procedure is repeated until the question has been fully answered or no relevant cognition structure is extractable or creatable.

Fig. 8 is a schematic illustration of a fourth processing operation according to the embodiment of the present invention. In more detail Fig. 8 shows an interaction between the client 4 and the server 5 and between the client 4 and the sensor 7. Since these interactions are a combination of the interactions shown in Figs. 12 and 13 a detailed description of these interactions is omitted here. However, it should be noted that these procedures can be performed simultaneously, successively or in a overlapping manner as the need arises.

Fig. 9 is a schematic illustration of a fifth processing operation according to the embodiment of the present invention. In more detail, Fig. 9 shows interactions between the client 4 and the sensor 7 and between the server 5 and the sensor 7. Since these interactions are a combination of the interactions shown in Figs. 13 and 14 a detailed description of these interactions is omitted here. However, it should be noted that these procedures can be performed simultaneously, successively or in a overlapping manner as the need arises.

Fig. 10 is a schematic illustration of a sixth processing operation according to the embodiment of the present invention. In more detail, Fig. 10 shows interactions between the client 4 and the server 5 and between the server 5 and the sensor 7. Since these interactions are a combination of the interactions shown in Figs. 12 and 14 the description of these interactions is omitted here. However, it should be noted that these procedures can be performed simultaneously, successively or in a overlapping manner as the need arises.

Finally, Fig. 11 is a schematic illustration of a seventh processing operation according to the embodiment of the present invention. In more detail, Fig. 11 shows interactions between the client 4 and the server 5, between the client 4 and the sensor 7 and between the server 5 and the sensor 7. Since these interactions are a combination of the interactions shown in Figs. 12 to 14 a detailed description of these interactions is omitted here. However, it should be noted that these procedures can be performed simultaneously, successively or in a overlapping manner as the need arises. Fig. 15 is a overall structure of a general system according to the embodiment of the present invention. As is obvious from Fig. 15 the embodiment of the present invention can be generalized such that several system nodes comprising clients 4, servers 5 and/or sensors 7 can be provided as system nodes 10 to 14 which are connected via a network 15. In order to constitute a distributed network comprising a plurality of system nodes 10 to 14 if the need arises.

It should be noted that there is generally described a transmission between system nodes because in a general system each of the system nodes can function as at least one of a client 4, a server 5 and a sensor 7. It is assumed that these system nodes are coupled via intelligent interfaces with each other for performing a communication. The intelligent interfaces transfer object features of cognition structure objects comprising knowledge, information and data depending on a respective question progressively more faithful to detail.

In the embodiment of the present invention there is also the possibility that specific cognition structure objects are stored locally at the side of one or more system nodes and can be used in further processing operations.

The aforementioned embodiment of the present invention is applicable to all kind of cognition structures such as data, information and/or knowledge comprising cognition structures based on physical quantities and cognition structures not based on physical quantities.

According to the embodiment of the present invention the capacity transmission channel can be extremely decreased by transmitting knowledge, information and/or data present in a hierarchical structure in the aforementioned manner.

Although the present invention has been described by way of a specific embodiment of the present invention it is not intended to delimit the present invention to the specific embodiment described above. Rather, the present invention is intended to include all modifications and amendments which fall within the scope of the present invention as defined in the appended claims.

Claims

Ansprϋche

1. Computer-implemented system for progressively transmitting of knowledge between system nodes (4, 5, 7) of a network structure, comprising: a plurality of system nodes (4, 5, 7); and intelligent interfaces by means of which respective system nodes (4, 5, 7) are coupled with each other for performing a communication therebetween, wherein the intelligent interfaces transmit object features of cognition structure objects comprising knowledge, information and data depending on a respective question of a respective one of the system nodes (4, 5, 7) progressively more faithful to detail from another of the respective system nodes (4, 5, 7) to the one of the respective system nodes (4, 5, 7).

2. System according to claim 1 , wherein relations are present between respective cognition structure objects.

3. System according to claim 1 or 2, wherein the cognition structure objects are hierarchically structured and data objects are located on a lowest level of the hierarchy.

4. System according to claim 3, wherein the progressive more faithful to detail transmitting is performed from a higher hierarchical level to lower hierarchical level within the hierarchy.

5. System according to one of the claims 1 to 4, wherein each system node comprises at least one of a client (4), server (5) and sensor (7).

6. System according to claim 4, wherein the server (5) comprises a database (6).

7. System according to one of the claims 1 to , wherein different kinds of cognition structure objects are arbitrarily processed leading to a multi-modal processing.

8. Computer-implemented method for progressively transmitting knowledge between system nodes (4, 5, 7) of a network structure, the method comprising the step of: progressively more faithful to detail transmitting object features of cognition structure objects comprising knowledge, information and data depending on a respective question of one of the respective system nodes (4, 5, 7) from another of the respective system nodes (4, 5, 7) to the one of the respective system nodes (4, 5, 7).

9. Method according to claim 8, wherein relations are present between respective cognition structure objects.

10. Method according to claim 8 or 9, wherein the cognition structure objects are hierarchically structured and data objects are located on a lowest level of the hierarchy.

11. Method according to claim 10, wherein the progressive more faithful to detail transmitting is performed from a higher hierarchical level to lower hierarchical level within the hierarchy.

12. Method according to one of the claims 8 to 11 , wherein each system node comprises at least one of a client (4), server (5) and sensor (7).

13. Method according to claim 12, wherein the server (5) comprises a database (6).

14. Method according to one of the claims 1 to 13, wherein different kinds of cognition structure objects are arbitrarily processed leading to a multi-modal processing.

15. Computer program product which forms directly or after performing a predetermined routine indirectly in cooperation with a computer or a computer system a system according to one of the claims 1 to 7

16. Computer program product which performs directly or after performing a predetermined routine indirectly in cooperation with a computer or a computer system a method according to one of the claims 8 to 14.