EP0962873A1
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Processing of textual information and automated apprehension of information
Assignee: International Business Machines Corp
•
Filing: 1998-06-02
•
Granted: —
•
Expires: 2018-06-02
Abstract
Abstract Scheme for the automated apprehension of textual information conveyed in an input string. The input string is segmented to generate segments and/or semantical units. The following steps are repeated for each segment in the input string until a subset for each segment in said input string is identified: a. identifying a matching semantical unit in a fractal hierarchical knowledge database of semantical units and pointers, said matching semantical units being deemed to be related to a segment of said input string, b. determining the fitness of said matching semantical unit by taking into consideration said semantical unit's associations, c. defining a subset of information related to said matching semantical unit within said fractal hierarchical knowledge database. Then these subsets are combined to form a resulting semantic network.
Inventors
- 👤 Gerd K. Binnig
- 👤 Peter Blöchl
- 👤 Jürgen Klenk
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-09 14:58
CPC Classifications
G; G06; G06F; G06F40/00; G06F40/20; G06F40/205; G06F40/211; G06F40/216; G06F40/253; G06F40/279; G06F40/289; G06F40/30; G06F40/35; Y; Y10; Y10S; Y10S707/00; Y10S707/99931; Y10S707/99933; Y10S707/99935
Claims Analysis
Total Claims: 55
Claim 1
Independent Claim
73 words
Method for the processing of textual information conveyed in an input string together
with information contained in a knowledge database representing a network of
hierarchically arranged semantical units which are similar across hierarchies, said
method comprising the steps: a. segmenting said input string into segments and b. combining said segments with semantical units from said knowledge database to
generate a resulting semantic network of hierarchically arranged semantical units
which are similar across hierarchies.
Claim 2
Dependent Claim
45 words
The method of claim 1, wherein at least one of said segments is deemed to be related
to a semantical unit, or is similar to a semantical unit, and wherein there are at least n
semantical units in the knowledge database with n ≥ 2.
Claim 3
Dependent Claim
34 words
The method of claim 2, whereby a semantical unit is a set containing one or several
pieces of information, such a semantical unit preferably consisting of a name and
pointers to other semantical units.
Claim 4
Dependent Claim
80 words
The method of claim 2, whereby step b. comprises the following steps i. identifying a matching semantical unit in said knowledge database, said matching
semantical unit being deemed to be related to the j th segment of said input string, ii. determining the fitness of said matching semantical unit by taking into consideration
said semantical unit's possible associations, repeating steps i. through ii. until a matching semantical unit is found for all j = 1,...,m
segments of said input string.
Claim 5
Dependent Claim
42 words
The method of claim 4, whereby the combining of said segments with said matching
semantical units is done such that the resulting semantic network comprises semantical
units that are deemed to be related to the textual information conveyed in said input
string.
Claim 6
Dependent Claim
27 words
Method of any of the preceding claims, wherein said resulting semantic network is
employed for the automated apprehension of the textual information conveyed in the
input string.
Claim 7
Dependent Claim
25 words
The method of claim 1, whereby said knowledge database and resulting semantic
network comprise several types of semantical units such as objects, relations, and
attributes.
Claim 8
Dependent Claim
18 words
The method of claim 1, whereby said knowledge database and resulting semantic
network comprise several types of pointers.
Claim 9
Dependent Claim
60 words
The method of claim 4, whereby steps i. through ii. are carried out m times if there are
m matching semantical units in said knowledge database which are deemed to be
related to the j th semantical unit of said input string, and whereby only one of said m
matching semantical units is used to generate said resulting semantic network.
Claim 10
Dependent Claim
44 words
The method of claim 4, whereby said matching semantical units are combined with
each other to form the resulting semantic network if a corresponding classification
probability indicates that the respective semantical units have been correctly matched
with the semantical units in said knowledge database.
Claim 11
Dependent Claim
21 words
The method of claim 4, whereby said semantical unit's possible associations are its
possible attributes and/or possible relations and/or possible roles.
Claim 12
Dependent Claim
53 words
The method of claim 4, whereby attributes and/or relations and/or roles of a
semantical unit in the input string are used to determine whether a matching semantical
unit in said knowledge database is better, meaning for example that it has a higher
fitness than any other matching semantical unit in said knowledge database.
Claim 13
Dependent Claim
30 words
The method of claim 4, whereby a bonus is added to increase the fitness if generally
accepted attributes and/or relations and/or rules are used in connection with a
semantical unit.
Claim 14
Dependent Claim
29 words
The method of claim 4, whereby a malus is subtracted to decrease the fitness if
unusual attributes and/or relations and/or roles are used in connection with a
semantical unit.
Claim 15
Dependent Claim
26 words
The method of claim 1 or 4, whereby each semantical unit in the knowledge database
receives a potential, said potential being a fixed or variable potential.
Claim 16
Dependent Claim
39 words
The method of claim 8, whereby a pointer between two semantical units in the
knowledge database has a weight corresponding to some kind of semantical distance
between the two semantical units, said weight being a fixed or variable weight.
Claim 17
Dependent Claim
18 words
The method of claim 1, wherein the resulting semantic network is a network identified
within said knowledge database.
Claim 18
Dependent Claim
22 words
The method of claim 1, wherein the resulting semantic network is extracted from the
knowledge database to form a new separate network.
Claim 19
Dependent Claim
19 words
The method of claim 2, wherein the input string is transformed into an input network
comprising said semantical units.
Claim 20
Dependent Claim
17 words
The method of claim 19, wherein the resulting semantic network is obtained from
transforming said input network.
Claim 21
Dependent Claim
17 words
The method of claim 8, wherein certain of the pointers describe the mutual
relationships between semantical units.
Claim 22
Dependent Claim
24 words
The method of claim 16, wherein the inverse of said weight represents some kind of a
semantical distance between two semantical units it connects.
Claim 23
Dependent Claim
26 words
The method of claim 4, wherein steps i. through ii. are repeated if an additional input
string is received until a resulting semantic network is generated.
Claim 24
Dependent Claim
15 words
The method of claim 1, wherein an input text is transformed into several input strings.
Claim 25
Dependent Claim
29 words
The method of claim 1, wherein a preliminary theme of the input string is determined
allowing a quick identification of a preliminary subset of information within the
knowledge database.
Claim 26
Dependent Claim
25 words
The method of claim 25, wherein said preliminary subset is from time-to-time,
or
continuously revised, preferably if a contradiction or change of theme is determined.
Claim 27
Dependent Claim
22 words
The method of claim 1, comprising the step of performing an action depending on the
general meaning extracted from said input string.
Claim 28
Dependent Claim
24 words
The method of claim 1, wherein an action is performed by providing an answer if the
input string was determined to comprise a question.
Claim 29
Dependent Claim
21 words
The method of claim 1, wherein the knowledge database comprises a self-similar
representation of semantical units and pointers across different scales.
Claim 30
Dependent Claim
17 words
The method of claim 4, wherein self-similar algorithms are used when carrying out
steps i. through ii.
Claim 31
Dependent Claim
25 words
The method of claim 1, wherein the knowledge database reflects the structure of an
environment, said environment preferably being a subset of the real world.
Claim 32
Dependent Claim
20 words
The method of claim 8, wherein at least one of said pointers is a directed associative
connection between semantical units.
Claim 33
Dependent Claim
36 words
The method of claim 8, wherein at least one of said pointers is a hierarchical pointer,
or a horizontal pointer, or a similarity pointer, or a functional pointer, or an
attributional pointer, or a role pointer.
Claim 34
Dependent Claim
17 words
The method of claim 1, wherein text is transformed into input strings which are then
processed string-by-string.
Claim 35
Dependent Claim
17 words
The method of claim 1, wherein speech is transformed into input strings which are
then processed string-by-string.
Claim 36
Dependent Claim
19 words
The method of claim 35, wherein speech recognition software is employed to
transform said speech into said input strings.
Claim 37
Dependent Claim
19 words
The method of claim 19, wherein the input network is generated by means of syntactic
parsing and/or grammatical parsing.
Claim 38
Dependent Claim
40 words
Method for the construction of a fractal hierarchical knowledge database, comprising
the steps: recording semantical units, linking said semantical units by means of pointers with other semantical units of the
fractal hierarchical knowledge database, assigning a weight to said pointer.
Claim 39
Dependent Claim
38 words
The method of claim 38, wherein said pointer is a hierarchical pointer or a horizontal
pointer, and said hierarchical or horizontal pointer is a similarity pointer, or a
functional pointer, or an attributional pointer, or a role pointer.
Claim 40
Dependent Claim
19 words
The method of claim 38, wherein there are various types of pointers which describe
the associations between semantical units.
Claim 41
Dependent Claim
14 words
The method of claim 38, wherein the pointers themselves are regarded as semantical
units.
Claim 42
Dependent Claim
16 words
The method of claim 38, wherein the weight assigned to the pointer is fixed or
variable.
Claim 43
Dependent Claim
38 words
The method of claim 38, wherein a fixed or variable potential is assigned to a
semantical unit, preferably when the fractal hierarchical knowledge database is used
for the automated apprehension of textual information conveyed in an input string.
Claim 44
Dependent Claim
19 words
The method of claim 43, wherein a fixed or variable threshold is assigned to said
fractal hierarchical knowledge database.
Claim 45
Dependent Claim
15 words
The method of claim 44, wherein the potential and threshold are used for subset
determination.
Claim 46
Dependent Claim
27 words
The method of claim 38, wherein the inverse of the weight of a pointer represents
some kind of a semantic distance between two semantical units it connects.
Claim 47
Dependent Claim
86 words
Apparatus for the processing of textual information conveyed in an input string
together with information contained in a knowledge database representing a network
of hierarchically arranged semantical units which are similar across hierarchies, said
apparatus comprising: means for segmenting said input string into segments, memory for storing said segments, a semantic processor for combining said segments with semantical units from said
knowledge database to generate a resulting semantic network of hierarchically
arranged semantical units which are similar across hierarchies. memory for storing said resulting semantic network.
Claim 48
Dependent Claim
40 words
The apparatus of claim 47, wherein at least one of said segments is deemed to be
related to a semantical unit, or identical with a semantical unit, and wherein there are
at least n semantical units with n ≥ 2.
Claim 49
Dependent Claim
77 words
The apparatus of claim 48, wherein said semantical processor comprises means for identifying a matching semantical unit in said knowledge database, said
matching semantical unit being deemed to be related to the j th segment of said
input string, such that a matching semantical unit is found for all j = 1,...,m
segments of said input string, and means for determining the fitness of said matching semantical unit by taking into
consideration said semantical unit's possible associations.
Claim 50
Dependent Claim
22 words
The apparatus of claim 49, comprising a speech processing unit, preferably comprising
speech recognition module, which transforms speech into said input string.
Claim 51
Dependent Claim
20 words
The apparatus of claim 47 being designed for the automated apprehension of the
textual information conveyed in the input string.
Claim 52
Dependent Claim
16 words
The apparatus of claim 51 comprising means which trigger a reaction depending on
the apprehended information.
Claim 53
Dependent Claim
19 words
The apparatus of claim 47 comprising an post-processor which transforms the
resulting semantic network to create an output string.
Claim 54
Dependent Claim
20 words
The apparatus of claim 47, wherein the resulting semantic network is represented as a
fractal hierarchical network of semantical units.
Claim 55
Dependent Claim
24 words
The apparatus of claim 48 comprising a semantic preprocessor for processing said input string to generate an input
network which comprises said semantical units.
References Cited (Backward Citations)
EP0689147A2
Cited
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.