ITTO20120303A1 - METHOD AND SYSTEM FOR CARRYING OUT ANALYSIS AND AUTOMATIC COMPARISON OF PATENTS AND TECHNICAL DESCRIPTIONS. - Google Patents

Description

Description of the Industrial Invention entitled: "METHOD AND SYSTEM FOR PERFORMING ANALYSIS AND AUTOMATIC COMPARISON OF PATENTS AND TECHNICAL DESCRIPTIONS"

DESCRIPTION

The present invention relates to a method and a system for carrying out automatic analyzes and comparisons of patents and technical descriptions.

The number of technologies commercially available in the technical field, but also in the field of consumer goods, is growing. Despite the progressive specialization of technologies, various resources are also dedicated to the diffusion of technical solutions outside their original field of application, which thus find a greater number of practical applications in new products and processes.

The opportunity to obtain both economic and strategic advantages by means of patents is often delayed, if not actually exploited, due to the difficulties linked to the number of existing patents to be analyzed and to the formal language used to describe the invention.

Every day thinkers, inventors and companies waste energy and time trying to evaluate the level of innovation, the patentability of an invention and the originality of the content.

For example, it is known that for about fifty years now well-tested theories have been proposed (TRIZ by Altshuller G., "Creativity as an Exact Science: The Theory of the Solution of Inventive Problems", Gordon & Breach Science pub., 1984 ) to compare and evaluate the inventive level of different inventions. These theories, to be applied, require the direct intervention of man and a considerable part of this work is dedicated to less intense applications of the human brain such as reading and building a mental image of the relationships existing between the components of the invention examined, rather than to understand the concepts expressed and the ontology of the invention.

There are several software packages that represent technical / technological contents based on their functional representation, an engineering subset of mathematical graph theory. Dedicated tools have proven to be really effective in speeding up the understanding of concepts, but despite these aids, large amounts of time are required to define a functional representation of an invention.

Functional analysis is a powerful tool for designers who do not want to limit themselves to introducing small optimizations of a mechanical system, but rather intend to introduce significant innovations in order to create truly new products. The functional description of a product is a description at an abstract level, such that different design solutions can be explored in the conceptual design phase, developing functional variants.

Furthermore, functional analysis can play an important role in patent circumvention: a careful functional analysis of a patent allows the identification of any ambiguous definitions such as to make it possible to overcome the patent constraints with minor modifications. Conversely, when writing a patent, a functional analysis of the text constitutes an effective verification of the quality of the work done.

Functional modeling is also often used at the level of detailed design: following the approach of the Suh (Suh N.P., "The Principles of Design", Oxford Press, 1990), the function is assumed as the goal and the project is broken down in functional requirements that are directly related to the design parameters at any level of abstraction. Since the products are defined and should be designed according to the function they must perform, various authors have classified the typical functions of mechanical design, with the aim of associating "subsystems" (assemblies of physical parts or components) and sub-functions in in such a way as to provide the designer with a set of modular units with which to develop the final product. An exhaustive description of these works is given by Kirschman (Kirschman C.F. et al, "Classifying Functions for Mechanical Design", Journal of Mechanical Design (ASME), Vol. 120, September 1998).

Several papers have been published to propose effective representations of the functions to describe the different aspects of the designer's intent, a crucial task for the development of aids to conceptual design; the purpose of these works is the definition of effective methods also for the representation of the relations existing between functions, such as the relations "decomposed-in", "conditioned-by", "increased-by", "described-as" (Shimomura et al., "Representation of Design Object Based on the Functional Evolution Process Mode", Journal of Mechanical Design (ASME), Vol. 120, June 1998).

In this context, inventors are developing tools, systems and methods to assist the designer in carrying out functional analyzes starting from different types of input data. In accordance with the definition of product architecture given by Ulrich (Ulrich, K.T., "The Role of a Product Architecture in the Manufacturing Firm" Research Policy, vol. 24, 1995), i.e. the way in which the designer puts in relationship functional elements with the respective physical implementations, much of the research activity in this field is addressed to the mapping of functions into components. The definition of general criteria to map the correspondence between assembly functions and features of a CAD system is obviously useful; this goal was also pursued by Line et al. (Line J.K. et al, "Calculation of Product Architecture Metrics Within a Solid Modeler", accepted for publication on Proceedings of the Institution of Mechanical Engineers Part B - Journal of Engineering Manufacture, 2001) albeit with a completely different strategy.

Nowadays it is possible to have different techniques and technologies to analyze texts in electronic format and simplify the reading of its content. In particular for scientific documents and patents these technologies have proved their effectiveness, however much work is still required to allow a fast and efficient comparison between similar documents, identifying the novelty elements in the document or sorting more documents by level of innovation.

Despite the enormous energies dedicated to the conception and development of software for the automatic analysis of electronic documents, the extraction and representation of information are usually carried out in the same field of application as the techniques and technologies already implemented: (1) technologies for the extraction of concepts most of the time neglect what already exists, leaving out the historical background and the rules inherent in the layout of the original document or in the state of the art of the technology; (2) the techniques / technologies for the representation of concepts are most often associated with visualization methods more related to the technology of extracting the concepts that for the very purpose of representing the results of the analysis carried out in a way that is usable to man.

In this context it is evident that the automatic analyzes of texts lead to the categorization of concepts by taxonomy or similar "text-based" techniques and that structured data such as programming languages or databases can be represented by the techniques of graph theory; both of these representations are natural in their field and show the actual advantages of the next evolutionary step in this field: hybrid texts and graph-based analyzes of electronic documents with textual and graphical representation of their contents.

A computer-based software system and a method in accordance with the principles of the present invention solve the problems described in the previous paragraphs thanks to the ability to perform automatic analysis and comparisons of patents and technical descriptions of engineering systems (such as technical reports, user manuals and maintenance, repair manuals etc.), in order to translate the text of the invention into a block diagram whose graphic representation is carried out in accordance with the main market standards, i.e. functional models, IDEF models, etc.

In fact, the proposed system, according to the principles of the present invention, is able to recognize the components of the system described in the analyzed document, even if they appear with different denominations. Furthermore, the hierarchy can be identified in terms of level of detail / abstraction of the identified components, as well as a classification into categories such as "assembly", "part", "portion". Finally, the system is able to analyze the functional interactions between the components and to identify the secondary products and the main product of the analyzed system.

The advantages of the invention just described and others besides which will be illustrated in the course of the description, are obtained by means of a method and a system according to claims 1 and 15 respectively. The preferential implementations and non-trivial variations of the present invention are claimed in the respective dependent claims.

These processing are carried out by the Text Analysis Modules which interact with each other and make use of the content of a customizable database of "stop words" and similar terms as described below. The results of this analysis can be enriched by making use of a semantic processor, such as that described in US-A-6,167,370, capable of analyzing an electronic document in order to extract the Subject, the Action and the Object from each sentence.

The results of the analysis of the Text Analysis Module can be provided to the user in different ways, in accordance with the functionality of the "Post Processing" Module of the present invention.

(1) The first output mode consists in providing the user with a graphical representation of the analyzed system, which allows a quick view of the system components, their hierarchical classification and their functional interactions.

(2) A comparison between the results of different analyzes can be performed using a set of quantitative indicators capable of assessing the relevance, complexity and level of innovation (according to Altshuller's definition) of the systems analyzed and (optionally) to follow their evolutionary path.

(3) The functions performed by or performed on a component in all the systems analyzed are recorded in a database in such a way as to be able to easily find all the possible functional uses of that component.

(4) The components capable of performing a given function, on the basis of the results of the analyzes performed, are registered in a database, in order to have a tool for finding possible technical solutions capable of performing a predetermined task.

The databases described in (3) and in (4) of the previous list can be easily integrated with CAD systems and tools to aid conceptual design.

The present invention will be better described by means of some preferential implementations, to be considered non-limiting examples, with reference to the attached figures in which: - FIG. 1 is a schematic representation of an example of implementation of the system in accordance with the principles of the present invention;

- FIG. 2 is a schematic representation of the main elements of the system architecture in accordance with the principles of the present invention;

FIG. 3 is a schematic representation of the Component Classification Module of the system in accordance with the principles of the present invention;

- FIG. 4 is a schematic representation of the method for identifying the components of the system in accordance with the principles of the present invention;

- FIG. 5 is a schematic representation of the method for classifying the components of the system in accordance with the principles of the present invention;

- FIG. 6 is an example of a graphic display of the information extracted from an analyzed invention;

- FIG. 7 is an example of a list of components with the relative levels of detail DL and the corresponding "supersystems" extracted from the description of the system illustrated in Fig. 6; - FIG. 8 is an example of the list of secondary products with the respective probability value of main product MPPV, evaluated according to the two approaches described in the detailed description of the present invention, extracted from the description of the system illustrated in Fig. 6;

- FIG. 9 is an example of the list of functional interactions between the identified components of the invention shown in Fig. 6;

FIG. 10 is an example of the development of the level of detail during the description of the invention shown in Fig. 6

An example of implementation of the present invention as illustrated in Fig. 1, includes: a CPU 11 which may comprise a "general purpose" personal computer, a network server or a minicomputer, with standard input / output peripherals such as a keyboard 12, a mouse 13, a scanner 14, a CD reader 17 and a printer 15. The system also includes standard communication ports towards public and / or private LANs, WANs, etc.

The user can enter the text to be analyzed by typing it directly from the keyboard 12, or through other input devices (such as the scanner 14 with OCR systems, the microphone 18 and "speech recognition" software etc.), or by selecting documents from local mass storage, local networks or the WWW network.

In accordance with the diagram of Fig. 2, an example of implementation of the present invention consists of a Temporary Deposit 20, a Text Analysis Module 30, a Customizable Database of "Stop Words" and Analogous Terms 40, a Database of Extracted Information 50 , a Post-Processing Module 60.

The text to be analyzed, entered by the user, is stored in the Temporary Repository 20. This text is processed by the Text Analysis module 30, making use of the Customizable Database of "Stop Words" and Analogous Terms 40 and (optionally) of a commercial semantic processor (external to the system described in the present invention), in accordance with the methodology described in the following paragraphs.

The results of the text analysis are then recorded in the Extracted Information Database 50.

The Post-Processing Module 60 provides the user with the information contained in the database 50, organized according to different formats, as described in the following paragraphs, for a quick and effective use of the analysis results. The output can be sent to any peripheral, to a local data storage unit or to a network.

The Component Identification Module 31 makes it possible to identify all the components of the system described in the examined text (ie in the case of a patent the components of the invention).

This task can be performed following two different techniques: search by "keywords" and semantic search (Fig. 4).

The search by keywords (Fig. 4, left), is based on the rule for writing patents (for example Title 37 of the U.S. Code of Federal Regulations), according to which each component of the system must be numbered in such a way that the same reference number is adopted throughout the entire description and in all the figures of the patent.

The following procedure is adopted:

1) search for numeric characters in the text;

2) the neighborhoods of words preceding and following each number are analyzed; the size of the neighborhood can be customized by the user with distinct values for the number of words that precede and those that follow the numeric characters; a default value of five words preceding / following the number is assumed; each neighborhood constitutes a row of the matrix of the component candidates;

3) filtering of the terms relating to "non-components": the rows containing words belonging to the set a) of the Database of "Stop Words" and Analogous Terms 40 are deleted;

4) recognition of synonyms and analogous terms among all lines containing the same numeric character: the list of synonyms is contained in the set b) of the Database of "Stop Words" and Analogous Terms 40;

5) identification of the intersection set of the words belonging to the lines containing the same numeric character; this set of words is assumed as the representative name of the component identified by the reference number present in the same row. If the first word following the numeric character is "of", the intersection set is evaluated only between the words to the left of the number, while the words to its right are used by the Component Classification Module 32. Steps 4) and 5) can be suitably reversed without any variation in the results.

The semantic search (Fig. 4 on the right) is based on the assumption that in the technical description of a system, the subject and the object of a sentence are in many cases components of the system itself, that is, in accordance with the theory of Altshuller, an elementary triplet TFA (Tool, the subject; Field, the action; Artifact, the object). If the Field of the TFA triad is an insignificant verb (from a functional point of view), the corresponding Tool and Artifact lose their role as candidate components of the system. The procedure consists of:

1) the text to be analyzed is processed with a commercial semantic processor (for example Cobrain ™, Knowledgist ™, Phrasys ™, Semantic Explorer ™, CREAX, Kiwilogic ™ etc.), thus extracting a TFA from each proposition;

2) filtering of the TFA (Tool-Field-Artifact) triples containing a Field belonging to the set f) of the Database of "Stop Words" and Analogous Terms 40;

3) collection of tools and artifacts that survived the previous filtering phase;

4) (optional) addition of a further set of candidate components making use of commonly available techniques for the identification of terms representative of the text content (such as statistical analysis, clustering, Bayesian networks, etc.);

5) among all the candidates making up the repetitions are clustered also taking into account the list of synonyms of the set b) of the Database of "Stop Words" and Analogous Terms 40.

The Component Identification activity can be carried out by following alternatively or in parallel the search by keyword and the semantic search. By combining these two techniques it is also possible to identify "supersystems": the Tools and Artifacts that do not belong to the set of components identified by a reference number are clearly components external to the system.

The product of the Component Identification Module is a list of denominations, representative of the conceptual components of the examined system; a reference number (the same reference number used in the text for the description of the system) is associated with each component name together with the list of all the alternative names adopted in the description of the system to identify that component (synonyms, pronouns, adjectives that they may have been filtered by the intersection action in step 5) of the keyword search, if they were sometimes omitted).

The results of the Component Identification Form 31 are recorded in the Extracted Information Database 50.

The Component Classification Module 32 sorts and classifies the components identified by the Component Identification Node 31 based on three main criteria: level of detail / abstraction; secondary products and main product of the invention; role in assembly. An example of implementation of the Component Classification Module 32 according to the principles of the present invention is illustrated in Fig. 3: it consists of three sub-modules, respectively the Hierarchy Identification Sub-Module 34, the Product Identification Sub-Module 35, the Sub-Module -Mechanical Implementation Analysis Module 36.

A schematic representation of the method for classifying the components of the system according to the principles of the present invention is shown in Fig. 5.

The procedure consists of the following steps:

1) analysis of descriptive phrases and specification expressions such as "... of ...": for example, as described in the Component Identification by keyword, sometimes the word following the numeric character representing the reference number of the component is the preposition "of", followed by another component (eg side N of the piston M; arm N of the support M etc.); in these cases it is assumed that the component preceding the preposition "of" is a "subsystem" of the component following the same preposition; in other words, the level of detail of the first component is at least one step higher than the level of detail of the second;

2) search for descriptive verbs: the list of descriptive verbs is contained in the set d) of the Database of "Stop Words" and Analogous Terms 40. A descriptive verb is a verb of the type "comprise", "to be made of", "to be constituted by" etc. If a neighborhood of words centered on the descriptive verb (the width of the neighborhood is a function of the verb, given that, depending on its meaning, one or more subjects and one or more objects are expected) contains components identified by the Component Identification Module 31, assumes that the components that precede the descriptive verb are subsystem / supersystem of the components that follow it, depending on the meaning of the verb itself: for example, from the phrase "the arm support comprises two guide rails, a screw drive and a carrier" one can infer that "arm support" is a Supersystem compared to "guide rails", "screw drive" and "carrier", ie it is at a higher level of abstraction.

This research is carried out also taking into account the alternative names of each component recorded by the Component Identification Form 31.

As in the case of Component Identification, the performances of the present invention can be increased by the use of a semantic processor, since a more reliable identification of the components that are actually the subject or object of the descriptive verb can be guaranteed: consider the previous example modified as follows: "the arm support comprises two guide rails, a screw drive arranged between the opposite guide rails, and a carrier"; in this case the phrase "arranged between the opposite guide rails" does not contain any object of the verb "comprise", but contains an identified component of the system ("guide rails").

3) Simplification of the hierarchy: it may happen that the hierarchy of levels of detail defined on the basis of the two previous steps present hierarchical jumps; for example, "A is a supersystem of B", "B is a supersystem of C" and in the text C is mentioned as a subsystem of A; by analyzing the complete list of hierarchical links between the components of the system, all jumps of this type are eliminated.

4) identification of parallel hierarchies: it may happen that one or more components are identified as subsystems of two or more components at the same level of detail; for example: A is a supersystem for Al and A2; B is a supersystem for B1 and B2; A2 and B1 are subsystems of C. These parallel hierarchies are appropriately registered since they usually correspond to the different functions performed by the same component.

A level of detail (DL) is assigned to each component in such a way that the Maximum level of abstraction is represented by DL = 0 and the DL of each subsystem is a higher level than the DL of the corresponding supersystem.

The results of the first classification phase can be further processed to identify the role of the components within the assembly described in the text: it is possible to assign to each component an "assembly", "part" or "portion" attribute with the aim of allow a more complete exchange of information with CAD / PDI / CRM systems; this information can be transferred using the most common data exchange formats, such as IGES, STEP, IDEF etc. 1) a "portion" attribute is assigned to all the components whose name contains words belonging to the set e) of the Database of "Stop Words" and Analogous Terms 40, such as "side of ...", "upper part of. .. "etc. In a CAD system these components correspond to geometric entities such as modeling features, faces, vertex edges etc.

2) an "assembly" attribute is assigned to all components that have at least one subsystem that has not been labeled as "leading" to the previous step;

3) a "part" attribute is assigned to all components not labeled in the previous two steps.

This classification allows a direct link with the Feature tree of a Part model and / or with the assembly tree of an assembly model in a CAD system, thus integrating the conceptual model of a mechanical system with its implementation.

A third way to classify the components of the examined system consists in identifying which of them are secondary products of the system itself and among these the main product. For this classification the results of the Interaction Analysis Module are required, and therefore will be described later.

The results of the Component Classification Form 32 are recorded in the Extracted Information Database 50.

The Interaction Analysis Module 33 allows the identification of the functional constraints existing between the identified components of the examined system. Also in this case two different approaches can be followed (keyword search and semantics):

1) search for word sequences containing the names of two system components separated by a verb not belonging to the list f) of the "Stop Words" Database and Analogous Terms 40; this verb is assumed to be the Field of a TFA triplet; it is assumed that the components preceding and following the verb are the Tool and the Artifact of the TFA triad, according to the meaning and form (active / passive) of the verb itself;

2) search for word sequences containing at least one component of the system and a verb belonging to the set c) of the Database of "Stop Words" and Analogous Terms 40 (verbs significant from the functional point of view): it is assumed that this verb is the Field of the TFA triad, while the component is considered Tool or Artifact according to its position with respect to the verb (previous or following), and the meaning and form (active or passive) of the verb itself. It is evident that the list c) of the Database of "Stop Words" and Analogous Terms 40 must contain all the forms that the verbs belonging to this list can take.

An alternative implementation of the same function makes use of a semantic processor: of each TFA triad, if both the Tool and the Artifact are components of the system and the Field is a verb not belonging to the list f) of the "Stop Words Database "and Analogous Terms 40, then this triad TFA is assumed as an elementary functional block of the system; otherwise, if only one of the Tool and Artifact is a system component, but the Field is a verb belonging to the set c) of the Database of "Stop Words" and Analogous Terms 40, it is assumed that the Tool / Artifact is not identified both an External Component of the system and the complete triad constitutes an elementary functional block of the system. Finally, it may happen that the Tool is a component of the system, but the Field-Artifact pair is itself a function of the system (for example, provide-tension = tighten); if a Field-Artifact pair among those extracted from the semantic processor belongs to the set g) of the Database of "Stop Words" and Analogous Terms 40, then the subject is considered the Tool of the triad and the Field-Artifact pair is translated into a Field functional in accordance with the table associated with the set g) of the Database of "Stop Words" and Analogous Terms 40. The search for the Artifact of this function can be left to the user or performed by searching for the first identified component that typically follows the preposition associated with this Field-Artifact pair (the prepositions most typically associated with Field-Artifact pairs are cited in the same table of the Database of "Stop Words" and Analogous Terms 40). For example, in the sentence "the upper end of the piston contacts the arm to provide tension on the arm which ..." the semantic processor identifies the triad "upper end - provide - tension" from which the "upper end" Tool are extracted , the Field "tighten" and the Artifact "arm" (as a component following the preposition "on".

At the end of the Interaction Analysis phase all the Tool / Artifacts not identified by the Component Identification Module 31 are considered External Systems and can be used to link or combine two or more inventions, technical systems and / or highlight the relationships between the examined system and the other systems belonging to the State of the Art of the field of application.

Usually, if the text to be examined is a patent, the Interaction Analysis is performed on the Detailed Description of the Invention and / or on the summary and / or on the claims.

At the end of the analysis of the interactions carried out by Module 33, all the TFA triples are recorded in the Database of Extracted Information 50, as well as the position in the text examined of the sentence from which this triplet was extracted: the position is evaluated by numbering sequentially all the sentences of the examined text, distinguishing the sentences from each other on the basis of the character "." 0 of the ASCII Carriage Return character.

By combining the results of the Component Identification Module 31, the Detail Hierarchy Identification Sub-Module 34 and the Functional Interaction Analysis Module 33, it is possible to identify all the secondary products and among them the main product of the analyzed system. In the example of implementation of the present invention, illustrated in Fig. 3, this task is performed by the Product Identification Sub-Module 35, in accordance with the following procedure:

1) each Artifact is a product of its Tool through the action of the Field of the TFA triad; therefore, in the first analysis, it is assumed that each Artifact is a secondary product of the examined system;

2) a secondary product loses this property (thus becoming a standard component of the system) in the following cases:

a. in the hierarchy in terms of detail / abstraction the candidate secondary product has at least two higher abstraction levels, i.e. its DL (Detail Level) is greater than or equal to 2;

b. the number of functional interactions such that the secondary product is a Tool is greater than or equal to the number of functional interactions in which it is an Artifact;

3) of each secondary product the ratio between the number of interactions such that it is an Artifact and the number of interactions in which it is a Tool is calculated; this ratio is called Principal Product Probability (MPPV);

4) alternatively, if the examined text is a patent, the MPPV factor can be calculated as the sum of the following partial probability values:

to. check if a secondary product is mentioned as an Artifact in the first two claims of the patent; example of partial probability value: 50%;

b. check if a secondary product is mentioned as an Artifact in the patent summary; example of partial probability value: 30%;

c. check if a secondary product is mentioned as an Artifact in the patent title; example of partial probability value: 15%;

d. calculation of the occurrences of secondary products in the entire text of the patent and normalization of these values with respect to the maximum; these normalized values are the partial probability value as a percentage, which in any case must not exceed a predefined value (eg 5%);

And. verifies whether a secondary product is an Artifact of a Field present as Field in the first two claims of the patent; example of partial probability value: 30%;

f. check if a secondary product is an Artifact of a Field present as Field in the patent summary; example of partial probability value: 20%;

g. check if a secondary product is an Artifact of a Field present as Field in the title of the patent; example of partial probability value: 10%;

h. calculation of the occurrences of the Fields that act on secondary products (in the role of Artifact) in the entire text of the patent and normalization of these values with respect to the maximum; these normalized values are a percentage of the partial probability value, which in any case must not exceed a predefined value (eg 10%);

i. calculation of the occurrences of the Field-Artifact pairs such that the Artifact is a secondary product in the entire text of the patent and normalization of these values with respect to the maximum; these normalized values are the partial probability value as a percentage, which in any case must not exceed a predefined value (eg 30%);

The Post Processing Module 60 provides the contents of the Extracted Information Database 50 to the user, organizing such data in four different forms, according to the purpose of the analysis. An example of implementation of the Post Processing Module 60, in accordance with the principles of the present invention, consists of four sub-modules as shown in Fig. 2: Graphical Mapping Module of the Text content 61; Text Comparison Module 62; Functional Uses Database of the examined components 63; Database of components capable of performing a predefined function 64.

The Text Content Graph Mapping Module 61 is an effective tool for communicating all the information extracted from the Text Analysis Module 30: the results of the Component Identification, Hierarchy Identification and Functional Interaction Analysis are represented in a diagram, such as in Fig. 6, with the following conventions:

1) each identified component of the system is represented by a rectangle labeled with its reference number and the representative name defined by the Component Identification Module 31; each component identified external to the system is represented by a hexagon labeled by means of an "EXT" string, a sequential number and the representative name defined by the Component Identification Module 31; the main product is represented by an ellipse labeled with the same criteria described above for the internal / external components of the system;

2) the hierarchy of levels of detail identified by the Classification Module 32 is represented by nesting the components at the highest level of detail within the corresponding components at the most abstract level;

3) the functional interactions between the identified components of the system are represented by arrows pointed by the Tool towards the Artifact, labeled using the Field, according to the results of the Interaction Analysis module 33;

4) upon request, some summary tables can be provided to the user:

to. list of components with the corresponding level of detail DL and the relative supersystem (Fig. 7);

b. list of secondary products as Field-Artifact pairs, with the corresponding MPPV Principal Product Probability Value (Fig. 8);

c. list of partial probability values evaluated according to the detailed procedure in the description of the Product Identification Sub-Module 35 (Fig. 8); d. list of functional interactions between the identified components (Fig. 9);

And. list of Field-Artifact pairs, such that the Artifact is the Main Product of the system.

The Text Comparison Module 62 allows the comparison between the descriptions of two or more systems, based on the following parameters:

1) comparison between the "diameters of the inventions", ie the number of levels of detail identified by the Component Classification module 32;

2) comparison between the number of internal components of the examined system, also distinguishing the number of components identified at each level of abstraction;

3) (if the analysis was carried out using the Mechanical Implementation Analysis Module 36) comparison between the number of "assembly", "part" and "portion" of the examined system; 4) comparison between the number of interactions identified by the Interaction Analysis Module 33; if two or more Fields are associated with the same Tool / Artifact pair, a check is carried out to eliminate the synonyms Fields, based on the list i) of the Database of "Stop Words" and Analogous Terms 40;

5) comparison between the number of interactions (counted as in step 4) that act on components at the same level of detail; possibility of highlighting whether these components belong to the same supersystem or not;

6) comparison between the number of interactions (counted as in step 4) that act on components at different levels of detail; possibility of highlighting whether these components belong to the same supersystem or not; the "hierarchical distance" between the interacting components is also recorded, ie the difference between their levels of detail DL;

7) comparison between the number and the lengths of the branches present in the functional diagram of the systems examined (such as the one in Fig. 6) evaluated starting from the Main Product of the system;

8) comparison between the components of the same rank: the rank of a component is defined as the minimum distance, in terms of the number of functional interactions, which bind the Main Product of the System to the component itself;

9) analysis of the trend of the level of detail during the description of the examined system (Fig. 10): the Interaction Analysis Module 33 records the position in the text of each of the TFA functional interactions identified; it is assumed that the level of detail of the Tool and the Art ifact in the sentence are the level of detail of the description, therefore it is possible to analyze the trend of the level of detail in the text and compare that of different texts. The analysis of the peaks of the Level of Detail during the description of a system allows the identification of the "core" of the invention, as well as of the secondary peculiarities.

The Database of the Functional Uses of the components in different systems 63 records all the functional interactions associated with homonymous components in all the examined texts, storing the reference to the source document and the role of the component in the TFA triad.

The user can be provided with a comparison between the different ways of using the same component in the different systems examined.

The Database of Components capable of performing a data 64 function records:

1) all the Tools associated with homonymous Fields found in all the examined texts, recording the reference to the source text and the complete TFA triad;

2) all the Tools associated with homonymous Field-Artifact pairs found in all the examined texts, recording the reference to the source text and the complete TFA triad;

The user can be provided with a comparison between the ways in which the same action (Field) or the same function (Field-Artifact) are performed in the different systems examined.

The Database of "StopWords" and Analogous Terms 40 consists of eight sets of words, all of which can be customized by the user either manually or through an automatic extraction procedure described below:

1) the user provides the system with a set of documents typical of the field of application in which he is interested; 2) a semantic analysis is performed by means of a commercial semantic processor (for example, Cobrain ™, Knowledgist ™, Phrasys ™, Semantic Explorer ™, CREAX, Kiwilogic ™ etc.) and a table of Tool and Artifact and of the their frequency;

3) starting from the comparison of the table defined in the previous step and the complete Database 40, it is possible to customize the lists of filters and synonyms, thus creating a typed subset of Database 40, labeled with the field of application of the documents processed in step 1) ·

Database 40 consists of the following sets:

a) list of "Stop Words" for words adjacent to numerical characters during the Component Identification activity; this set is typically made up of references to figures, patents or other documents, units of measurement etc.

b) table of synonyms of the candidate components, at different levels of detail (for example, portino, side, end; piston, plunger etc.);

c) list of fields typical of the functions required in a given field of application;

d) table of descriptive verbs such as "to comprise", "to be made of", "to be constituted by", etc. This list must take into account all the forms that these verbs can take, also due to conjugation irregularities; for example, the forms "comprise", "comprises", "comprised", "comprising" etc. are recorded together with the infinitive "to comprise".

e) lists of terms that describe a portion of a component, such as "end", "side", "face", "part" etc.

f) list of verbs that are not significant from a functional point of view;

g) table of Field-Artifact pairs, their translation into a functional verb and one or more typically associated prepositions, used for the automatic search of the Artifact; for example, "provide tension", "tighten", "on";

h) table of synonyms of functional verbs representing a Field.

The "Stop Words" and Analogous Terms Database 40 allows customization of the system described in the present invention, but other customizations are also possible:

1) the activities of the Component Identification Module 31, Component Classification Module 32 and Interaction Analysis Module 33 can be monitored step by step by the user who can compare the information extracted with the relative source sentence, or they can be performed automatically, albeit with lower reliability;

2) the user can specify a list of components (Tools / Artifacts) and / or functions (Fields) on which to focus the analysis of the interactions, so that only the relative functional sub-diagram is extracted;

3) the search for secondary products and / or the main product of the systems examined can be limited to components external to the system.

Further characteristics of the present invention are given by:

* the Post Processing module 60, in which the Text Content Module 61 represents:

a. each identified component of the system with its reference number and representative name defined by the Component Identification Form 31;

b. any identified component or element external to the system;

c. the main product;

d. the hierarchy of levels of detail determined by the Classification Module 32;

And. the functional interactions between the components identified according to the results of the Interaction Analysis Module 33;

* the Text Content module 61 which represents:

- each component identified in the system by means of a rectangle labeled with its reference number and the representative name defined by the Component Identification Form 31;

- each identified component or element external to the system is represented by a hexagon labeled with the string "EXT", a sequential number and the representative name defined by the Component Identification Module 31;

- the main product by means of an ellipse labeled with the same criteria illustrated above for the internal / external components;

- the hierarchy of the levels of detail determined by the Classification Module 32 represented by nesting the components at a level of detail more pushed internally to the components at a higher level of abstraction;

- the functional interactions between the identified components are represented by arrows pointing from the Tool to the Artifact, labeled with the Field, in accordance with the results of the Interaction Analysis Module 33;

* the Post Processing module 60 in which the Text Content Module 61 represents:

to. a list of components with their DL level of detail and the corresponding supersystem;

b. a list of secondary products in the form of Field-Artifact pairs with the corresponding MPPV main product probability value;

c. a list of partial probability values evaluated in accordance with the detailed procedure in the description of the Product Identification sub-module 35;

d. a list of functional interactions between the identified components;

* the Post Processing module 60 in which the Text Comparison Module 62 allows the comparison between two or more system descriptions according to the following parameters:

to. comparison between the "diameter" of the systems, defined as the number of levels of detail identified by the Component Classification Module 32;

b. comparison between the number of internal components of the systems examined, taking into account both the complete list of identified components and those at each level of detail;

c. (if the analysis of the Mechanical Implementation Analysis sub-module 36 has been carried out) comparison between the number of "assemblies" "parts" and "portions" of the examined system;

d. comparison between the number of interactions identified by the Interaction Analysis module 33; if two or more Fields are associated with the same Tool-Artifact pair, a check is carried out to eliminate synonyms Fields, taking into account the set i) of the database of "stop Words" and Analogous Terms 40;

And. comparison between the number of interactions (counted as in step d) acting on components at the same level of detail; it is possible to highlight whether these components are internal or external to the system;

f. comparison between the number of interactions (counted as in step d) acting on components at different levels of detail; it can be highlighted whether these components are one subsystem of the other or not; it is also possible to take into account the "hierarchical distance" between the interacting components, ie the difference between their levels of detail;

g. comparison between the number and length of the branches present in the functional diagram of the systems examined (such as the one in Fig. 6) evaluated starting from the Main Product of the systems themselves;

h. comparison of components of the same rank: the rank of a component is defined as the minimum distance in terms of number of interactions, which connects the Main Product of the system with the component itself;

i. analysis of the trend of the level of detail during the description of the examined system: the Interaction Analysis Module 33 records the position in the text of each TFA interaction; the level of detail of the Tool and of the Artifact in a sentence is assumed as the level of detail of the description, therefore it is possible to analyze the trend of the level of detail in the examined text and compare this trend in different texts;

* the Post Processing module 60 in which the analysis of the peaks of the trend of the level of detail during the description of the system allows the identification of the core and the secondary peculiarities of the system itself;

* the Post Processing module 60 in which the database of the Functional uses of the Components in the different systems 63 records all the functional interactions associated with homonymous components in all the examined texts, recording the reference to the source text and the role of the component in the TFA triad;

* the Post Processing module 60 in which the database of Components Capable of Performing a Data 64 Function records: a. all the tools associated with homonymous Fields in all the texts examined, recording the reference to the source text and the complete triad;

b. all the Tools associated with the homonymous Field-Artifact pairs found in the examined texts, recording the reference to the source text and the complete triad;

* the database of "Stop Words" and Analogous Terms 40 which consists of eight sets of words, all customizable by the user through the following automatic extraction procedure:

a) the user provides the system with a set of documents typical of the field of application he is interested in;

b) a semantic analysis is performed by means of a commercially available semantic processor (for example, Cobrain ™, Knowledgist ™, Phrasys ™, Semantic Explorer ™, Kiwilogic ™ etc.) and a table of Tool and Artifact is recorded and the number of occurrences;

c) comparing the table defined in the previous step and the complete database 40 it is possible to automatically customize the lists of filters and synonyms, thus creating typed sub-sets of the database 40, labeled with the field of application of the documents processed in step 1);

* the database of "Stop Words" and Analogous Terms 40 which is made up of the following sets:

a) list of stop keywords for words adjacent to numerical characters during the Component Identification activity; this set is typically made up of references to figures, patents or other documents, units of measurement etc. b) table of synonyms of the candidate candidates, at different levels of detail (for example, portion, side, end; piston, plunger etc.);

c) list of fields typical of the functions required in a given field of application;

d) table of descriptive verbs such as "to comprise", "to be made of", "to be constituted by", etc., taking into account in this list all the forms that these verbs can take, also due to irregularities of the conjugation;

e) list of terms describing a portion of a component, such as "end", "side", "face", "part" etc. f) list of verbs that are not significant from a functional point of view;

g) table of Field-Artifact pairs, of their translations into a functional verb and one or more prepositions typically associated with this term used for the automatic search of the Artifact; h) table of synonyms of functional verbs representing a Field;

* the system in which customization of the following systems is allowed:

to. the activity of the Component Identification Module 31, Component Classification Module 32 and Interaction Analysis Module 33 can be followed step by step by the user, who can compare the extracted information with the corresponding source sentence, or it can be carried out automatically although with less reliability ; b. the user can specify a list of components (Tool, Artifact) and / or functions (Field) on which to focus the analysis of functional interactions, in order to extract the corresponding functional sub-diagrams; c. the search for secondary products and the main product of the systems examined can be limited to components external to the system.

Claims (12)

CLAIMS 1. Method suitable to be automatically executed in a computer program to carry out automatic analysis and comparisons of patents and technical descriptions of engineering systems through a hierarchy of components in terms of level of detail, said method comprising the steps of: - examine a text of a patent or a technical description; identify system components described in the examined text; classify the role of said components identified in terms of categories such as "assembly", "part", "port ion"; classifying said components identified in a hierarchy in terms of level of detail / abstraction; recognize all functional links and interactions existing between said identified components of the examined system; translating said identified components into a block diagram; compare components and descriptions between two or more systems; And - organize the data in different forms according to the purpose of the analysis. 2. Method according to claim 1, characterized in that the identification of all the components of the system described in the examined text is performed according to the following procedure: to. search for numeric characters in the text; b. for each number, analysis of a neighborhood of the preceding and following words, each neighborhood constituting a row of the matrix of the component candidates; c. filtering of the "non-component" terms, deleting the lines containing the words adjacent to the numeric character; d. recognition of synonyms and similar terms between lines containing the same numeric character; And. identification of the intersection set of the words belonging to the lines containing the same numeric character, assuming this set of words as representative of the component identified by the numeric character of these lines. 3. Method according to claim 1 characterized by the fact that the identification of all the components of the system described in the examined text is performed on the basis of the assumption that said components interact as subjects and objects of an elementary functional triad TFA (Tool, subject; Field, action; Artifact, object) in accordance with the following procedure: to. extraction from each sentence of a triad TFA (Tool-Field-Artif act), for example from an XML document or by means of a semantic processor; b. filtering of the TFA triples containing a list of verbs that are not meaningful from a functional point of view; c. collection of the Tools and Artifacts that survived the previous filtering phase; d. optionally, adding a further set of component candidates making use of commonly available techniques for identifying words representative of the content of a text; And. among all component candidates (Tool and Artifact surviving the filtering phase), removal of repeated names, also taking into account the synonyms of the component candidates. 4. Method according to claim 1, characterized in that a detail / abstraction comparison criterion is applied to classify the components of the system in accordance with the following steps: a. analysis of descriptive phrases and specification expressions such as "... of ..."; b. assignment to the components preceding the preposition "of" of the subsystem role of the system following the preposition "of"; c. search for descriptive verbs such as "to comprise", "to be made of", "to be constituted by" etc., taking into account all the forms that these verbs can take, also due to the irregularities of the conjugation; d. assignment to the previous components of a verb describing the role of subsystem / supersystem of the components following the same descriptive verb according to the meaning of this verb; And. carrying out this analysis taking into account all the alternative names of each component, in which a Level of Detail (DL) is assigned to each component, representing said DL the maximum level of abstraction when DL = 0, and such that each subsystem is a higher level of the corresponding supersystem, and, if different Levels of Detail (DL) are assigned to the same component, such that the maximum level of abstraction is represented by DL = 0 and the DL of each subsystem is a level greater than the DL of the corresponding supersystem, a simplification of the hierarchy is carried out by eliminating the hierarchical jumps, and if different Levels of Detail (DL) are assigned to the same component, such that the maximum level of abstraction is represented by DL = 0 and the DL of each subsystem is a level greater than the DL of the corresponding supersystem, the identification of parallel hierarchies takes place taking into account such "parallel" hierarchies. 5. Method according to claim 1, characterized in that all components are further processed in order to identify the role of a component in the assembly described in the text, in accordance with the following procedure: to. a "portion" attribute is assigned to all components whose name contains terms that describe the portion of a component, such as "end", "side", "face", "part", etc .; b. an "assembly" attribute is assigned to all components that have at least one subsystem that has not been labeled as "portion" in the previous step; c. a "part" attribute is assigned to all components not labeled in the previous two steps. 6. Method according to claim 3, characterized in that an identification of the functional links existing between the identified components of the examined system is performed in accordance with the following steps: to. search for word sequences containing the names of two system components separated by a verb, excluding the component-verb-component triples such that the verb is not meaningful "from the functional point of view"; b. attribution to the components that precede and follow said verb of the role of Tool and Artifact of the triad, according to the meaning and the active or passive form of the verb itself; c. search for word sequences containing at least one component of the system and a verb of the required functionalities in a given field of application, that is, the verbs that are significant from a functional point of view; d. attribution to said component of the role of component of the triad, according to the meaning and the active or passive form of the verb itself; in which an external system is identified, this system being a Tool or an Artifact of a TFA functional triad, and where, if a functional link is identified and such that the Tool is a component of the system and the Field-Artifact pair can be translated into a function, the search for the Artifact of this function is delegated to the user or performed by searching for the first component identified following the preposition typically associated with this Field-Artifact pair. 7. Method according to claim 6, characterized in that all secondary products of the examined system are identified in accordance with the following procedure: to. each Artifact is assumed to be a secondary product of the examined system; b. a secondary product loses its characteristic, thus becoming a "standard" component of the system, in the following cases: - in the hierarchy of the level of detail a secondary product candidate has at least two levels of superior abstraction, that is, its Level of Detail or Detail Level, DL, is greater than or equal to 2; - the number of functional interactions such that the candidate secondary product is a Tool is greater than or equal to the number of functional interactions such that it is an Artifact. 8. Method according to claim 1, characterized in that the main product of the examined system is identified among all the secondary products, as the one for which the sum of the following different probability values is maximum: to. verifying if a secondary product is mentioned as an Artifact in the first two claims of the patent; b. verifying if a secondary product is mentioned as an Artifact in the patent summary; c. verifying if a secondary product is mentioned as an Artifact in the patent title; d. evaluating how many times the secondary products are mentioned in the entire patent and normalizing these values with respect to the maximum frequency, this normalized value multiplied by 100 being assumed as a partial probability value, but in any case having to be less than or equal to a predefined value; And. verifying if a secondary product is an Artifact of a Field present as Field in the first two claims of the patent; to. verifying if a secondary product is an Artifact of a Field present as Field in the patent summary; b. verifying if a secondary product is an Artifact of a Field present as Field in the title of the patent; c. evaluating how many times the Fields acting on secondary products, considered as Artifact, are mentioned in the entire patent and normalizing these values with respect to the maximum frequency, this normalized value multiplied by 100 being assumed as a partial probability value, but in any case it must be lower or equal to a predefined value; d. by evaluating how many times the Field-Artifact pairs, such that the Artifact is a secondary product, are mentioned in the entire patent and normalizing these values with respect to the maximum frequency, this normalized value multiplied by 100 being taken as a partial probability value, but in any case it must be less than or equal to a predefined value. 9. System for carrying out automatic analyzes and comparisons of patents and technical descriptions of engineering systems according to the method of claim 1, said system being characterized in that it comprises: - a Temporary Archive 20 in which the text to be analyzed entered by the user is recorded; - a Database of Stop Words and Analogous Terms 40; - a Text Analysis Module 30 from which the text is processed; - a Database of Extracted Information 50; - a Post-Processing Module 60; - a Component Identification Form 31 designed to identify all the components of the system described in the examined text; - a Component Classification Form 32 designed to sort and classify the identified components; - an Interaction Analysis Module 33 designed to identify the functional links existing between the identified components of the examined system, all the links identified being capable of being registered in the Extracted Information Database 50; - a Product Identification sub-module designed to identify all secondary products and among them the main product of the system examined; - the Post Processing Module 60 being able to supply the content of the database of the extracted information 50 to the user, organizing such data in different forms according to the purpose of the analysis. 10. System according to claim 9, characterized by the fact that it further comprises a semantic processor suitable for carrying out the identification of all the system components described in the examined text, said semantic processor being able to extract from each sentence a triad TFA (Tool- Field-Artifact), called semantic processor comprising: a. means for filtering the TFA (Tool-Field-Artifact) triples containing a Field belonging to a set of the database of Stop Words and Analogous Terms 40 which contains a list of verbs that are not significant from a functional point of view; b. means to collect the surviving tools and artifacts of previous filtering means; c. optionally, means for adding a further set of component candidates by means of commonly available techniques for identifying words representative of the text content, ie statistical analyzes, clusterizers, Bayesian networks; d. among all the component candidates, said components being Tool and Artifact surviving the filtering means, means to eliminate the repeated names also taking into account the list of synonyms of a set of the database of Stop Words and Analogous Terms 40 containing a table of synonyms of candidate members, at a different level of detail; And. means for assuming all remaining components as components of the examined system. 11. System according to claim 9, characterized in that it further comprises a semantic processor which is able to identify the functional links existing between the recognized components of the examined system, said semantic processor being able to extract from each sentence a triad TFA (Tool- Field-Artifact) and including: a. if both the Tool and the Artifact are components of the system and the Field does not belong to a set of the database of Stop Words and Analogous Terms 40 which contains a list of verbs that are not meaningful from a functional point of view, means to consider this triad TFA an elementary functional block of the system; b. otherwise, if only one of the Tool and Artifact is a component of the system, but the Field is a verb of the required functionalities in a given field of application, means to consider that the missing Tool / Artifact is an External Component of the system and to consider the triad completes an elementary functional block of the system; c. if a Field-Artifact pair among those extracted from the semantic processor belongs to a set of the database of Stop Words and Analogous Terms 40, their translations into a functional verb and one or more prepositions typically associated with this phrase, used to automatically search for the 'Artifact, means to consider the subject of the verb as a Tool of the triad and to translate the Field-Artifact pair, in accordance with the table of the database of Stop Words and Analogous Terms 40, into a functional Field. 12. System according to claim 9, characterized in that attributes such as "assembly", "part" or "portion" identifying the role of a component in the assembly described in the text are suitable for being linked to the geometric database of a CAD system with a direct association to the feature tree of a part model and / or to the assembly tree of an assembly model, thus integrating a conceptual model of a mechanical system with its implementation.