7

 [7]
The clustering center point (a COGS) 371.50 of the alternate sense, “*lincoln*” expression 371.5 is a point or small subregion in the space of the primitive cognitions layer 371 of keyword space 370 to which that alternate sense expression, “*lincoln*” (371.5) is anchored. Unlike the keyword node 371.5 (Kw1?=“*lincoln*”, but in another cognitive sense), the clustering center point (COGS) 371.50 is not given a specific name or other articulable attributes by system users. Instead, this data object (the COGS 371.50) operates like a shadowy entity that represents a cognitive sense, where the represented COGnitive Sense (where the capitalized letters explain where the acronym COGS comes from) is inferred from the keyword nodes closest to it and where the distances (hierarchically and/or spatially speaking) of the clustered-about nodes relative to the given, cognitive-sense-representing clustering center point (e.g., COGS 371.50) indicates how close in a cognitive sense way, the cognitive senses of the respective nodes are to that of the center point (e.g., COGS 371.50). In other words, the first mentioned Kw1 of the given example, “*lincoln*” (371.1) represents “*lincoln*” taken according to a respective, first cognitive sense (e.g., the 16th President of the United States or 16th POTUS) while the second mentioned Kw1? of the given example, “*lincoln*” (371.5) represents “*lincoln*” taken according to a respective, second and different cognitive sense (e.g., the Lincoln™ brand of automobiles or the city of Lincoln, Nebr.) and the shadowy, cognitive-sense-representing clustering center points (e.g., 371.0, 371.50) which are most closely disposed (hierarchically and/or spatially) to the respective keyword nodes that have a same keyword expression (e.g., “*lincoln*”) but different cognitive senses for the same, respectively represent the cognitive sense but without providing an “expression” (e.g., Lincoln, the 16th President; or Lincoln, the automobile brand) for that cognitive sense. Instead, each of cognitive-sense-representing clustering center points 371.0 and 371.50 respectively draws its represented cognitive sense from the keyword expressing nodes (e.g., Kw1, Kw2, Kw1? Kw6) closest to it. It is essentially a symbiotic relationship. The one or more closest COGS (e.g., 371.50) adjacent to a given keyword node gives a cognitive sense form of spin to the keyword expression (e.g., “*lincoln*”) of that node while the one or more closest keyword nodes (e.g., Kw1? Kw6) to a given COGS (e.g., 371.50) inferentially give cognitive sense to that COGS (e.g., 371.50). If system users vote to add-to or delete or move the keyword nodes (e.g., Kw1? Kw6) that are closest to a given COGS (e.g., 371.50), such a user-driven change can alter the inferred cognitive sense of the corresponding COGS. On the other hand, if system users vote to add or delete or move the closest COGS's that surround a given keyword node (e.g., Kw2), such a user-driven change can alter the cognitive sense spin that is projected onto the keyword expression (e.g., “*lincoln*”) of that node by the nearest cognitive-sense-representing clustering center points (COGS's).
The hierarchical and/or spatial space of the primitive cognitions layer 371 shown in FIG. 3E can be 2-dimensional, 3-dimensional or of greater dimensionality and/or it can have a hierarchical organization wherein PNOS-type points, nodes or subregions thereof are linked in accordance with a hierarchical tree structure. In one embodiment, hierarchical and/or spatial distance away from a given clustering center point (COGS, e.g., 371.50) indicates how dissimilar, far away, or unlike the inferred cognitive sense of the clustering center point 371.50 is the cognitive sense of each expression (e.g., Kw6?*) 371.6 that is disposed in that primitive cognitions layer 371 of the keyword space 370. In other words, other keyword expressions that are anchored relatively close to, or at zero distance from the given clustering center point 371.50 are respectively deemed to be correspondingly similar to, or same as, in a cognitive sense of the other keyword expressions (e.g., Kw6?*, 371.6) while those that are calculated to be farther away (hierarchically and/or spatially) are deemed to be proportionally more distant or dissimilar in terms of their respective cognitive senses.
In terms of a more concrete example, assume that the cognitive sense of, as well as the expressional equivalent of the alternate sense expression, “*lincoln*” (371.5) is “Lincoln, Nebr.; the City of”. Assume that the cognitive sense of, as well as the expressional equivalent of the nearby Kw6 expression node 371.6 is “Nebraska; The Capital City of”. It turns out that Lincoln Nebr. is the Capital City of the State of Nebraska. Therefore, although the expressions “Lincoln, Nebr.; the City of” and “Nebraska; The Capital City of” are not the same expressions, under a cognitive sense analysis they refer to substantially the same cognitive concept. Hence the hierarchical and/or spatial distance between points, nodes or subregions 371.5 and 371.6 should be approximately zero. In one embodiment, a relative pointer 371.56 that logically links node 371.5 (Kw1?) to node 371.6 (Kw6) includes an indication of how far away, hierarchically and/or spatially, from starting position 371.50 (the clustering center point) is the nearby node 371.6 (Kw6). In this case, the first exemplary node 371.5 (Kw1?) is assumed to be positioned dead center on top of clustering position 371.50 (the clustering center point or COGS). The nearby other node 371.6 (Kw6) is deemed to be slightly spaced apart, and in a corresponding direction, from the clustering center position 371.50 (a relative origin). The data that represents relative pointer 371.56 may also include an indication of the location in system memory where the nearby expression Kw6 (371.5) is stored as well as hierarchical and/or spatial vector indicating how far away and in what direction the nearby expression Kw6 (371.5) is displaced relative to the center point expression Kw1? (371.5).
In similar fashion, the first used example of keyword expression Kw1 (node 371.1), where its expression, “*lincoln*” is determined by communal consensus to refer to the Abraham Lincoln sense of that expression, is located dead center over different clustering center point 371.0. A relative distancing and direction pointer 370.12 (which like other pointers discussed herein is understood to be a stored physical signal pointing to a stored other physical signal, e.g., the one representing second keyword Kw2) is provided to indicate that the second keyword expression Kw2 has a substantially same or similar cognitive sense as does the first keyword expression Kw1 even if the second keyword expression Kw2 is substantially different from the first keyword expression (e.g., “16th USA President” versus “Ab* Lincoln”). (Because illustration space is relatively tight in FIG. 3E, some concepts relating to cognitive sense center points, e.g., COGS's 371.0 and 371.50 and to vectors pointing away therefrom (e.g., 371.56 or 370.12) and to other kinds of pointers (371.52) will be discussed while referring to one rather than the other. However, it is to be understood that the generic aspects of the descriptions apply to both.)
As indicated by the above, each respective, clustering center point (e.g., COGS's 371.0 and 317.50—each represented in FIG. 3E by a cross hatched ellipse) may provide a Thesaurus like or semantic type of contextual flavor to the various expressions (e.g., Kw1, Kw1?) that are positioned either directly over the respective center points and to the other keyword expressions that are hierarchically and/or spatially disposed as spaced apart but clustered nearby and around the respective clustering center point (e.g., 371.0 or 317.50). It is left up to respective governance bodies (a.k.a. herein as relevant communities) who are in charge of the different subregions of the context space to determine what Thesaurus like or semantic type or other cognitive sense is applied by the respective clustering center point (e.g., 371.0 or 317.50) and this is done by how they position the various nodes nearby to the given COGS. More specifically, these cognitive senses are implicitly defined when the hierarchical and/or spatial positions of the consensus-wise created clustering center points (e.g., 371.0 or 317.50) are created by, or revised by corresponding controlling communities of users (a.k.a. governance bodies) and when the various keyword expressions (e.g., Kw1, Kw1?) are positioned either directly over or nearby the respective center points (COGS's) and/or when so-called, operator nodes (see 372.1) are operatively coupled to the primitive layer expressions having the different cognitive senses and/or when so-called, operator nodes (see 372.1) are operatively positioned (hierarchically and/or spatially) adjacent to their own nearby cognitive-sense-representing clustering center points (COGS's, not shown for the illustrated operator nodes due to space limitations in the drawings).
As mentioned, each consensus-wise created or communally-updated clustering center point (e.g., 371.0 or 317.50) has assigned to it a respective hierarchical and/or spatial position in the space of the corresponding Cognitions-representing Space or subregion thereof (e.g., keyword expressions primitives layer 370). Each clustering center point (e.g., 371.0 or 317.50) also has assigned to it a first creation date indicating time stamp and optionally, a list of later position and/or cognitive sense modification dates. Each clustering center point (e.g., 371.0 or 317.50) further has assigned to it a primary expression pointer (not shown) that points to the one keyword expression (e.g., Kw1 371.1) that is deemed by the controlling community to be the expression which is most closely linked with the respective clustering center point (e.g., 371.0). Each clustering center point (e.g., 371.0 or 317.50) may further have assigned to it, one or both of re-direction and expansion pointers 371.52 (both represented by the one arrowed line in FIG. 3E, see also 30W.7ERR of FIG. 3W).
After a clustering center point (e.g., 371.0 or 317.50) is first created by a corresponding governance body and the hierarchical and/or spatial areas around it are populated by associated keyword expressions (e.g., Kw2, Kw3, Kw4, etc.) it may become desirable to add yet further keyword expressions in same close proximity with the cognitive sense represented by the first created center point (e.g., 371.0). However, it may become inconvenient or impractical or otherwise not proper to crowd all the new keyword expressions around the same center point (e.g., 371.0). Instead, it may become desirable to create a “twin” (e.g., 371.51) of the first created center point (e.g., 371.50) in another location of memory. This may be done with use of a so-called, center point “expansion” pointer 371.52 (see also 30W.7ERR of FIG. 3W). The latter points bi-directionally as between the earlier created original (e.g., 371.50) and later-in-time created twin (e.g., 371.51) and also provides a date stamp as to when the twin was created. Keyword expressions that attach to the later-in-time created twin (e.g., 371.51) inherit the creation date of that twin rather than the creation date of the original center point (e.g., 371.50). Therefore it becomes possible with this data structure to determine the timing of the cognitive sense that is attached to a given newer keyword expression as opposed to the perhaps slightly different, cognitive sense that is attached to an earlier created keyword expression. Also, legacy hierarchical and/or spatial assignments may be preserved.
Alternatively or additionally after a clustering center point (e.g., 371.0 or 317.50) is first created by a corresponding governance body and the hierarchical and/or spatial areas around it are populated by associated keyword expressions (e.g., Kw2, Kw3, Kw4, etc.) it may become desirable to drastically change the keyword expressions associated with that earlier-in-time center point (e.g., 371.50) and/or to drastically change the hierarchical and/or spatial distancings between the surrounding keyword expressions and the center point (e.g., 371.50). At the same time, it may be desirable to preserve legacy structures. Accordingly, rather than erasing an originally created structuring of clustering center points (e.g., 371.0 or 317.50) and surrounding expression nodes thereof (e.g., Kw1 and Kw1?), a re-directing pointer (represented by the same link 371.52 as used for the expansion pointer, see also 30W.7ERR of FIG. 3W) may be attached to each originally created center point (e.g., 371.50) and that time stamped, “re-directing pointer” 371.52 is understood by the system software to mean, don't use this center point but rather jump to the next (newer) center point (e.g., 371.5) and use that next (newer) center point as if it were this center point. Re-directing pointers can of course be cascaded to form a linked list that redirects a software action originally directed to an original center point to instead be applied to a substitute center point created many levels later. In this way the system can adapt to ever changing cognitive senses and sentiments (over time and/or user populations) of its evolving user base. One of the redirected software actions may be one where the software is accessing keyword expressions located hierarchically and/or spatially a given distance away from and/or in a given direction away from the originally specified center point (COGS). In other words, if the software is instructed to fetch all keyword expression nodes disposed within X distance from the identified center point (e.g., COGS 371.50) and redirection is in effect, the software will instead fetch all keyword expression nodes disposed within X distance from the alternate center point (e.g., 371.51) to which it was redirected by pointer 371.52. This allows legacy software to transparently access that latest (most up to date) communally created and communally updated version of keyword space (KWs 370) even though the legacy software code tells it to reference the earlier in time and originally created keyword center point (e.g., 370.50). Incidentally, although the data objects representing cognitive-sense-representing clustering center points (COGS's) do not have textual expressions defining their respective cognitive senses, they do each have a unique center point identifying field (not shown, see instead 30T.1 b of FIG. 3Ta as being an equivalent) so that the COGS's can be uniquely identified even if they have moved about hierarchically and/or spatially within their respective Cognitions-representing Space (e.g., keyword expressions space 370 of FIG. 3E). As a result, the system has an adaptively updateable, expressions, codings, or other symbols clustering layer (e.g., 371) that may be transparently updated by means of expansion and/or re-direction without having to change the legacy software that references it.
In one embodiment, each primary keyword expression node (e.g., Kw1 371.1) of a respective first clustering center point includes a linked list pointer pointing to the next node having a same or substantially same keyword expression but located in a different clustering area. For example, linked list pointer 371.49 may link from the node (371.1) of expression Kw1 to the node (371.5) of identical expression Kw1? (node 371.5 which is located over different clustering center point 371.50). The latter node would have a similar linked list pointer (not shown) pointing to the next node also having the same keyword expression (e.g., “*lincoln*”) but a different cognitive sense represented by a respective other clustering center point (not shown). In one embodiment, the linked list pointers (e.g., 371.49) also each include a pair of expression ranking values that rank the expressions at the terminal ends of the respective linked list pointer according to which is the most popular cognitive sense for that expression and which is the least. For example, the expression, “911” may have earlier had the cognitive sense of an emergency phone number as its number one ranked sense, However, after September 2001 the World Trade Center attack becomes the new number one ranked sense, System software can quickly scan through the linked list of pointers to find the current, top N cognitive senses for a given expression, where N can be 1, 2, 3, . . . here.
While clustering center points (e.g., 371.0, 371.50, 371.51) have been described thus far as providing, in one instance, a Thesaurus like or semantic type of flavoring to the keyword(s) overlaid directly on top of, or disposed hierarchically and/or spatially nearby to the respective clustering center point (e.g., keyword nodes 371.1, 371.5 and 371.6), more generally speaking, clustering center points (COGS's) can be used to imply other kinds of cognitive senses to respective PNOS-type points, nodes or subregions of other types of Cognitions-representing Spaces (e.g., music space, emotion space, historical events space) where there is no easy way (or any way) to articulate a communal “sense” that a relevant community cognitively attributes to the PNOS's that are disposed hierarchically and/or spatially in close proximity with each respective clustering center point (COGS). More specifically and for example, certain ones of advertising jingles or popular show tunes or movie scenes may evoke in a relevant community (e.g., a specific demographic group) a particular cognitive sensation that cannot be easily described with words and yet, when two or more of those advertising jingles or popular show tunes or movie clips are played to that demographic audience as representative examples of the cognitive sense, the audience knows it when it hears it (or knows it when they see it, this referring to the played video clips). Yet more specifically, and in the case of an American audience, the showing of a first image depicting the raising of the American flag at Iwo Jima, a second image depicting George Washington crossing the Delaware River and a third image depicting George W. Bush with a bullhorn at the attacked World Trade Center site soon after 9/11 may evoke certain emotions of patriotic pride and yet that cognitive sense cannot be easily put into words. In accordance with the present disclosure, nodes representing images such as these (and/or movie clips of this kind) may be closely clustered in a respective imagery space (see for example primitive data object 30M.0 of FIG. 3M) over or substantially close to a respective clustering center point (COGS) that directly represents the unarticulated cognitive sense (e.g., one associated with patriotic American pride as a nonlimiting example).
An example use for such a clustering center point is as follows. Assume that a user of the STAN—3 system recalls the imagery of the raising of the American flag at Iwo Jima (World War II) as one example that evokes patriotic pride and the crossing of the Delaware as a second “of its kind”, but the user does not remember yet further examples and the user wants to identify such further examples as understood by a given sub-community among system users. To this end the user instructs the STAN—3 system to find for him (or her) a closely clustered group of images in a system-maintained Image-type Cognitions-representing Space where two of the closely clustered representations of images (imagery nodes) are the ones for the recalled cases of Iwo Jima and the crossing of the Delaware. In response, the system automatically searches the given Cognitions-representing Space (and/or other interrelated spaces) for one or more clustering center points (COGS's) that have two such images in close proximity thereto, or overlaid directly on that found one or more clustering center points. More specifically, such a found clustering center point may additionally have adjacent thereto, images of specific events taking place at the Arlington Cemetery, or in front of the Lincoln Memorial, or with the Statue of Liberty as a backdrop, and so on. In other words, the system can automatically find others “of its kind” (as defined by respective user sub-communities) once a cognitive sense is hinted at by two or more user-provided examples that fit under the vague specification of, find for me more “of its kind” like these two or more examples. Stated otherwise, a given user sub-community may communally cross-associate in its communal mind, certain imageries, songs, historical events, etc. that belong together because they satisfy a perhaps-unarticulable cognitive sense (e.g., a communal “common sense”). The here-disclosed clustering center points enable the clustering together of such communally cross-associated items about respective clustering center points (COGS's) even if there is no one clear topic or central keyword expression or other specifiable other node that can tie the loose ends together just as well.
In one embodiment, the postionings in system memory of clustering center points are defined by absolute (long form) address pointers (e.g., stored in a lookup table (LUT) that cross-associates the COGS unique ID with its memory storage address and its hierarchical and/or spatial positioning) while the postionings in system memory of keyword nodes (e.g., 371.1, 371.2) clustered around that center point are defined by relative (short form) address pointers that use the center point address as a base. As a result, the bit lengths of digital pointers (memory address references) that point to the keyword primitives can be made relatively short while just one long-form base address is used for pointing to the corresponding clustering center point (e.g., 371.0).
Alternatively or additionally after a clustering center point (e.g., 371.0 or 317.50) is first created by a corresponding governance body and the hierarchical and/or spatial areas around it are populated by associated keyword expressions (e.g., Kw2, Kw3, Kw4, etc.), thereby defining relative distances between the various keyword nodes, it may become desirable to alter those represented distances. However, locations in hierarchical and/or spatial space are already defined for the originally created and center point surrounding nodes. In one aspect of the present disclosure, rather than changing the defined locations in hierarchical and/or spatial space of the already formed nodes, an altered distance calculating file or record is added to the definition of the clustering center point. The altered distance calculating file or record is represented by symbol 371.56 (but see also 30W.7ERR of FIG. 3W) and it may call for calculating of effective distances in various linear or nonlinear and/or condition based ways. Such altered distance calculations may include the use of one or more lookup tables (LUT's). Accordingly, if a legacy software module is instructed to access keyword expressions located hierarchically and/or spatially a given distance away from and/or in a given direction away from an originally specified center point, the distance (and angular direction) recalculating file/record is automatically consulted and is used to redefine the distance that are calculated (and/or looked up via LUT's) for respective keyword nodes. In other words, if the software is instructed to fetch all keyword expression nodes disposed within distance X from the identified center point (e.g., 371.50) or from another point whose position is specified relative to the identified center point and the distance recalculation/look-up functionality is in effect, then the software will instead fetch all keyword expression nodes disposed within a different X? (prime) distance, where that primed distance is computed (e.g., obtained with aid of lookup tables) according to the alternate distance calculating scheme (e.g., 371.56) attached to the specified clustering center point. This allows legacy software to transparently access the latest (most up to date) communally defined version of keyword space (KWs 370) per communally re-defined spacings between keyword-expression holding nodes (this includes operator nodes like 372.1) even though the legacy software code tells it to use a distance specified earlier in time and per the originally positioned keyword nodes. As a result, the system has an adaptively updateable, expressions, codings, or other symbols clustering layer (e.g., 371) that may be transparently updated without having to change the legacy software that references it or the originally specified positionings of the keyword expression holding nodes.
Assume for sake of a more concrete example of how primitives may be combined by operator nodes that the illustrated second keyword node 371.2 is disposed in the primitives holding layer 371 fairly close, in terms of spatial and/or hierarchical clustering (and optionally also in terms of memory address number) to the location assigned to the first keyword expression-holding node 371.1. Assume moreover, that the keyword expression (Kw2) of the second node 371.2 covers the expression, “*Abe” and by so doing (with asterisk in front) it covers the permutations of “Honest Abe”, “President Abe” and perhaps many other such variations. As a result, the Boolean combination calling for Kw1 AND Kw2 may be found in many of so-called, “operator nodes” for representing cognitions such as those related to “Honest President Abe Lincoln” and the like. An operator node, as the term is used herein, is provided and functions somewhat similarly to an ordinary expression-containing node in a hierarchical tree structure (and it inherits some attributes of its base or parent node(s)—see FIG. 3Q) except that it generally does not store directly within it, all the definitions of its intended, combined-primitive attributes. More specifically, if a first operator node 372.1—which node is shown disposed in a sequences/combinations layer 372 of FIG. 3E—were an ordinary primitive node rather than an operator node, that primitive node would directly store within it, the textual expression, “*lincoln* AND *Abe” (if the Abe Lincoln example is continued here). However, in accordance with one aspect of the present disclosure, operator node 372.1 contains references to one or more predefined functional “operators” (e.g., AND, OR, NOT, parenthesis, Nearby(number of words), After, Before, NotNearby( ), NotBefore, and so on) and it contains pointers as substitutes for variables that are to be operated on by the referenced functional “operators”. One of the pointers (e.g., 370.1) can include a long or absolute or base pointer having a relatively large number of bits and pointing to a predefined, clustering center point 371.0 while another of the pointers (e.g., 370.12) can be a short or relative or offset pointer having a substantially smaller number of bits because it uses the clustering center point 371.0 as a base for its represented offset value. This scheme allows the memory space consumed by various combinations of primitives (two primitives, three primitives, four, . . . 10, 100, etc.) to be made relatively small in cases where the plural ones of the pointed-to primitives (e.g., Kw1 and Kw2) are clustered together (spatially, hierarchically and/or address-wise) in the primitives holding layer (e.g., 371) around a same clustering center point (e.g., 371.0). In other words, rather than using two long-form pointers, 370.1 and 370.2 (the latter being shown for purpose of comparison, offset 370.12 is preferably used instead) to define the “AND”ed combination of Kw1 and Kw2, the first operator node 372.1 may contain just one long-form pointer, 370.1, and associated therewith, one or more short-form pointers (e.g., offset 370.12) that point to the same clustering region of the primitives holding layer (e.g., 371) but use the one long-form pointer (e.g., 370.1) as a base or reference point for addressing the corresponding other primitive object (e.g., Kw2 371.2) with a fewer number of bits because the other primitive object (e.g., Kw2 node 371.2) is typically clustered in a Thesaurus like or semantic contextual like clustering way around a clustering center point to which one or more keystone primitives (e.g., Kw1 node 371.1) are directly tied. In one embodiment, the relative offset pointer 370.12 (but see also 371.56) functions as a distance indicator because its offset from the clustering center point 371.0 can also represent distance in hierarchical and/or spatial space from the clustering center point.
While FIG. 3E shows pointers such as 370.1, 370.4, 370.5 etc. pointing upwardly in the hierarchical tree structure, it is to be understood that the illustrated hierarchical tree structure is navigatable in hierarchical down, up and/or sideways directions such that children nodes can be traced to and from their respective parent nodes, such that parent nodes can be traced to and from their respective child nodes and/or such that sibling nodes can be traced to and from their co-sibling nodes. In the illustrated example, operator node 372.1 is a child of the two parent nodes, 371.1 (Kw1) and 371.2 (Kw2) from which it inherits at least some of its internalized data. Pointers 370.1 and 370.2 point backwards to indicate the sources of the inherited, and thus incorporated by such reference data. However, from a hierarchical tree perspective, operator node 372.1 is the child of its two parent nodes, 371.1 (Kw1) and 371.2 (Kw2).
It is stated above that, often, keyword expressions (e.g., Kw1 371.1 and Kw2 371.2) come to be clustered together spatially and/or hierarchically next to one another and near a clustering center point (e.g., 371.0). But the mechanisms that can cause this close clustering together of nodes to happen have not been fully explained above yet. One option is that the spatial (e.g., in keyword space) and/or hierarchical (e.g., within a keyword ‘A’-tree) clustering together of semantically belonging-together keyword expressions is initially established on a permanent or modifiable basis by manual intervention by system operators and/or by trusted system users who have been granted privileges to manually assign spatial and/or hierarchical locations to all or a pre-specified subset of initial points, nodes or subregions of one or more Cognitive Attention Receiving Spaces (e.g., keyword expressions space). In that case, the so-privileged system operators/trusted users may organize the spatial and/or hierarchical placements of cognition-representing primitive and some higher level data-objects (e.g., keyword expressions) such that those that sensibly belong together are clustered together. More specifically, system operators and/or trusted system users may initially populate a primitives layer of a textual cognition space (e.g., keyword space, URL space, etc.) with multiple and spaced apart copies of textual expression clustering center points (e.g., 370.1, 371.50, etc.) paired directly with respective textual expression nodes (see FIG. 3W) containing the textual expression, “*lincoln*” where a first of such operator created pairing of a clustering center point and its directly overlying keyword node is assigned to the Abraham Lincoln sense of “*lincoln*”; where a second of such operator created, pairing of a clustering center point and overlying keyword node is assigned to the Lincoln, Nebr. sense of “*lincoln*”; where a third of such operator created, pairing of a clustering center point and overlying keyword node is assigned to the Lincoln Car Dealerships sense and so on.
Alternatively or additionally, the spatial and/or hierarchical placements of cognition-representing data-objects such as the keyword expression representing ones (e.g., 371.1, 371.2, 373.1), URL expression representing ones (e.g., 391.2, 394.1), meta-tag expression representing ones (not explicitly shown—see 395) are voted on by one or more direct or indirect voting mechanisms, where the vote is for continued approval of a current placement or for moving to a newly proposed placement, and/or continued approval of the current way the expression is expressed or for changing to a newly proposed way of expressing it (with characters or other symbols or codes). In response to such voting, the STAN—3 system automatically and responsively modifies the spatial and/or hierarchical placements of cognition-representing data-objects and/or of their contained expressions according to results of such voting mechanisms. One example of indirect (implicit) voting is when, as a result of a chat or other forum participation session, a subset of keyword expressions (e.g., 371.1, 371.2, 373.1) are determined to be the top N keywords now most popular with participants of the forum; in which case the popularity-wise clustered set of keyword expressions may be given corresponding nudges towards becoming clustered closer together (not necessarily over a clustering center point such as 371.0) in terms of their spatial and/or hierarchical placements within the corresponding Cognitive Attention Receiving Space (e.g., keyword expressions space). If enough chat or other forum participation sessions give cumulative nudges in a same direction to one or more such keyword expression holding nodes (e.g., 371.1, 371.2), the system responds by moving them closer together in the spatial and/or hierarchical placement sense. In accordance with one aspect of the present disclosure, some keyword expression holding nodes (e.g., 371.1) may be assigned a greater anchoring strength at their current position than others. As a result, when certain keywords are determined to have increased commonality with each other such that they merit being nudged closer together, the one with the greatest anchoring strength moves the least and the others therefore move toward its original location in hierarchical and/or spatial space. (The concept of anchoring will be discussed at greater length below in conjunction with 30R.9 d of FIG. 3R.)
While co-popularity among all users (or among a pre-specified subset of users; e.g., expert users) is one basis for nudging together into closer co-clustering with one another and in a corresponding hierarchical and/or spatial space of certain keyword expressing nodes (e.g., 371.1, 371.2, 373.1 as one example, but could be other nudged together points, nodes or subregions in other Cognitive Attention Receiving Spaces as a more general example), it is within the contemplation of the present disclosure to have oppositely acting mechanisms that nudge apart (and thus de-cluster in a spatial or hierarchical sense) certain groups of cognition representing data objects one from another. A more specific example will be given by way of section 30T.12 e 8 of FIG. 3Tb (to be described). For sake of a simple example here, let it be assumed that one user in one chat room has proposed that the keyword expression. “Goldwater” should be clustered together with the keyword expressions for Abe-Lincoln and Gettysburg Address. Let it be assumed that essentially all other involved users voted strongly (e.g., with great emotional intensity) against the idea. In other words, they were indicating that the keyword expression, “Goldwater” is greatly disliked (despised, negatively viewed) among a super-majority (e.g., 67% or more) of involved users and thus they were voting for nudging the keyword expression, “Goldwater” far away in spatial and/or hierarchical space from the keyword expressions that overlie the co-related cognitions of Abe-Lincoln and Gettysburg Address. (Clustering center points such as 371.0, 371.50 and 371.51 are the data objects that implicitly represent the underlying cognitive sentiments of their directly overlying expression nodes, although those underlying cognitive sentiments do not have to be explicitly spelled out. They can be implied by the placement of their directly overlying and/or further spaced away, expression-holding nodes.) As a consequence of a placement proposal and votes for or against it, if enough users (e.g., a number greater than a predetermined threshold) vote negatively against the proposal and/or if enough highly-influential experts (who may be given greater voting weights) vote implicitly or explicitly in such a negative or de-clustering direction, then the system will respond by automatically moving the keyword expression node for “Goldwater” (not shown in FIG. 3E) farther away in the spatial and/or hierarchical placement sense from the other clustered together data objects (e.g., 371.1, 371.2) which better represent the cognitive concepts of Abe-Lincoln and Gettysburg Address (as an example, see also briefly, node 30W.14 of FIG. 3W). With repeated votes of these pull-together and/or push-apart kinds and as recognized over pre-specified time spans (or all of system time) and/or as cast by different and optionally differently weighted users and/or users fitting pre-specified filtering criteria (e.g., demographic criteria in terms of age, gender, income level, geographic location etc.), the various points, nodes or subregions (e.g., keyword expressions) are jostled about in the respective keyword expressions space (or other corresponding cognition-representing space) until some come to be clustered closely together relative to one another and others come to be de-clustered and thus spaced relatively farther apart in the spatial and/or hierarchical sense. In accordance with one aspect of the present disclosure, a same cognition specifying data object (e.g., keyword expression, and more specifically, as an example from above, the expression, “*lincoln*”) can be repeated many times within a corresponding Cognitive Attention Receiving Space (e.g., keyword expressions space) where each instance has a respective different sense such that, in one instance, it is clustered closely together with a second such data object (e.g., “Goldwater” being clustered closely together with “Abe-Lincoln”) and in another instance it is spaced far apart in a spatial and/or hierarchical sense from the same second such data object because the cognitive senses of the different instances of the same expression are different. Each topic node may point to a respective, clustered together set of keyword expressions or the like (other clustered together cognition representing data objects) as is appropriate for that topic node and the users who favor that topic node. Two topic nodes may appear to correspond to a same topic and yet the users who favor the respective first and second topic nodes may have entirely different viewpoints regarding which other cognition representing data objects (e.g., top N keywords) are to be most liked (e.g., most popular) and which, if any, are to be most disliked (e.g., most despised, most emotionally rejected). In other words, the system allows for a wide variety of differing points of view as among different communities of system users. An data-objects organizing system for allowing such a thing to happen will be explained in yet more detail when FIG. 3R is described below.
Automatic clustering and/or de-clustering of the cognition representing data objects (e.g., topic nodes, keyword expression nodes, etc.) within the spatial and/or hierarchical space of a corresponding Cognitive Attention Receiving Space (CARS, e.g., topic space, keyword expressions space, etc.) need not be limited to or based only on the above described indirect voting where; as a result of a chat or other forum participation session, a subset of cognition representing data objects (e.g., keyword expressions 371.1, 371.2, 373.1 in FIG. 3E) are determined to be the top N such data objects (e.g., keywords) which are most popular in a positive favoring sense among participants of a corresponding forum or topic node and are thus urged to be spatially and/or hierarchically clustered closer together (e.g., in keyword space) and/or where a cognition representing data object (e.g., keyword expression 371.6) is determined to be among a top N? despised data objects (e.g., keywords) which are most unpopular (despised, viewed in a negative or disfavoring sense) among participants of the corresponding forum (or topic node) and is thus urged to be spatially and/or hierarchically moved apart from the favored, clustered together other such data objects (e.g., in keyword space). Various expert, credentialed and/or reputable or otherwise well regarded users may be given cluster-altering empowerments whereby their positive and/or negative, implicit or explicit votes operate to automatically urge movement of concurrently co-liked data objects closer together (tighter clustering) in a given data-objects organizing space and/or to automatically urge movement of a disliked data object further apart in spatial and/or hierarchical space from other nearby data objects that are voted upon by the empowered user as being “liked” for its/their current placement in spatial and/or hierarchical aspect of the given data-objects organizing space. In one embodiment, participants of chat or other forum participation sessions that tether strongly to a given vicinity (e.g., predefined subregion) of a Cognitive Attention Receiving Space are asked to vote on which among them is to act as a cluster-controlling representative who will be empowered to vote on behalf of the others (as a community representative) with regard to how corresponding cognition representing data objects should clustered close together or not within a given vicinity of a given data-objects organizing space (e.g., keyword space) in which the community is interested in.
Referring next to FIG. 3Q, shown here is an exemplary but not limiting (and not fully detailed) data structure 30Q.0 for defining an operator node. Due to space limitations in the drawings, some details of data structure 30Q.0 are left out, including for example, a set of linked list pointers similar to 30W.7 b of FIG. 3W and one or more pointers similar to 30W.7 c of FIG. 3W that point to a corresponding one or more nearest clustering center points. The below discussion re 30W.7 b and 30W.7 c of FIG. 3W are incorporated by reference here as if applied to the illustrated operator node data structure 30Q.0. In the illustrated example of FIG. 3Q, a first field 30Q.1 indicates the size, shape, location (e.g., relative location in a corresponding space, for example keyword space), identification and/or assigned virtual mass (or anchoring strength) of the operator node object. (As noted above, the data structure 30Q.0 may also or alternatively indicate an anchoring strength in place of or in addition to the virtual mass of the represented cognition-representing data object.) As also already mentioned above, an operator node uses pointers to draw into its definition, data from more primitive other data objects (e.g., from primitive cognition representing data objects and/or from functioning-as-parents, other operator nodes). When serving as part of a respective spatial (and optionally also hierarchical) space, the operator node may be assigned a respective virtual shape, a respective virtual size (e.g., virtual range of occupancy in a corresponding space), a respective virtual center of gravity, and optionally a respective virtual mass or virtual anchoring strength for that respective spatial space. The operator node should also have a unique identification code to distinguish it from other operator nodes of the same space. Often the operator node may be pictured as a movable spherical node of constant radius and having its mass temporarily rooted at a single point (center of gravity point) within the respective spatial space that is under consideration. Referring briefly to the perspective diagram of FIG. 3R, the three equally-sized spheres illustrated as residing inside of cylindrical space 30R.10 may alternatively represent operator nodes in place of the sibling topic nodes 30R.9 a, 9 b and 9 c that they do represent. However, spatial space in which such nodes are virtually placed is not limited herein to the 3-dimensional kind and operator nodes are not limited to ones that can be pictured as same sized and same shaped virtual objects (e.g., spheres) residing at respective locations within a given spatial (and optionally also hierarchical) space. More to the point, it is within the contemplation of the present disclosure to allow for the representing of any respective one of points, nodes or subregions within a respective spatial space (e.g., a 3-dimensional cylindrical kind) by means of a corresponding one or more operator nodes in place of a primitive node. So in the general sense, an operator node can be assigned a respective virtual shape and virtual size; particularly if it is to define a corresponding subregion within its given space (e.g., a subregion containing a plurality of virtual points). Additionally, the operator node will be assigned a unique virtual location where that assigned virtual location may (in one embodiment) coincide with the center of gravity of its assigned shape and mass. That assigned virtual location may (in one embodiment) coincide with the assigned location of a clustering center point (e.g., 371.0 of FIG. 3E) provided within the respective Cognitions-representing Space (e.g., keyword space). Accordingly, the first field 30Q.1 (the size/shape/location field) may contain data indicating the assigned virtual sizes, virtual shapes and/or virtual locations of the uniquely identified (ID'ed) operator node object within respective virtual spaces. Virtual distances between operator nodes whose virtual locations are adjacent to one another may indicate how closely clustered or not those operator nodes are to one another and/or to nearby clustering center points (e.g., 371.0 of FIG. 3E). In one embodiment, closely clustered operator nodes (and/or closely clustered cognition primitives) lend anchoring support to one another when nudges are applied for separating them from one another. This concept will be better explained when anchor 30R.9 d of FIG. 3R is described below. The point is that cognition representing data objects (e.g., operator nodes) that are deemed to be alike to one another, or otherwise as “belonging together”, may be automatically urged into clustering with one another in a hierarchical and/or spatial sense and the latter has implications when the respective virtual space is explored by a user or a search bot to see which points, nodes or subregions are clustered closely to one another (and thus deemed to be substantially same or similar to one another) and which are spaced far apart (and thus deemed to be substantially dissimilar to one another in terms of one or more cognitive senses covered by nearby clustering center points).
In one embodiment, the virtual size/shape/location field 30Q.1 may additionally provide real world information about the memory space consumed by data structure 30Q.0 (e.g., in terms of number of bits or words) and/or information about how the remaining fields of data structure 30Q.0 are organized.
A second field 30Q.2 of data structure 30Q.0 lists pointer types (e.g., long, short, operator or operand, etc.) and numbers and/or orders in the represented expression of each. A third field 30Q.3 contains a pointer to an expression structure definition that defines the structure of the subsequent combination of operator pointers and operand pointers. The operator pointers logically link to corresponding operator definitions. The operand pointers logically link to corresponding operand definitions. An example of an operand definition can be one of the keyword expressions (e.g., 371.6) of FIG. 3E. An example of an operator definition might be: “AND together the next N operands”. More specifically, the illustrated pointer to Operator definition #2 might indicate: OR together the next M operands (as pointed to by their respective pointers: Ptr. to Operand#2a, Ptr. to Operand#2b, etc.) and then logically AND the result with the preceding expression portion (e.g., Operator#1=NOT and Operand#1=“Car?”). The organization of operators and operands can be defined by an organization defining object pointed to by the third field. As mentioned, this is merely a nonlimiting example.
Aside from including operand and operator indicators (e.g., 30Q.5, 30Q.4), the data structure 30Q.0 of the operator node will typically include one or more, so-called, inheritance fields 30Q.H by way of which the data structure 30Q.0 inherits data structure parts of base level primitives and/or of its parent nodes of the one or more Cognitive Attention Receiving Spaces (CARSs) the operator belongs to. More specifically, most primitives will include a field containing pointers to points, nodes or subregions in the same and/or other Cognitive Attention Receiving Spaces (e.g., nodes or subregions of topic space) and/or a field containing pointers to chat or other forum participation sessions or other informational resources. The operator node (30Q.0) will similarly, and by means of inheritance (30Q.H) contain such pointers as well so that the operator node (30Q.0) can function as cross-linking data object just as can the base level primitives of the CARSs to which its operand pointers (e.g., 30Q.5) point to and/or so that the operator node (30Q.0) can function as a cross-referencing data object to informational resources just as can the base level primitives of the CARSs to which it belongs.
Referring back to FIG. 3E, in accordance with another aspect of the present disclosure, primitive defining nodes (e.g., Kw2 node 371.2) may include logical links to semantic or other equivalents thereof (e.g., to synonyms, to homonyms) and/or logical links to effective opposites thereof (e.g., to antonyms). A pointer in FIG. 3Q that points to an operand may be of a type that indicates an optional attribute such as: include synonyms and/or include homonyms and/or include or swap-in the effective opposites thereof (e.g., to antonyms). Thus, by pointing to just one keyword expression node (e.g., 371.2 of FIG. 3E) an operator node object (e.g., 372.1) may automatically inherit synonyms and/or homonyms and/or antonyms of the pointed-to one keyword (e.g., 371.2). The concept of incorporating effective equivalents and/or effective opposites applies to other types of primitives besides just keyword expression primitives. More specifically, a URL expression primitive (e.g., 391.2) might be of a form such as: “www.*lincoln*” and it might further have a logical link to another URL primitive (not shown) that references web sites whose URL's satisfy the criteria: “www.*honest?abe*”. Thus, a URL's combining operator node (e.g., 394.1 in FIG. 3E) might inherency-wise make reference to web sites whose URL name includes, “Honest Abe” (as an example) as well as those whose URL name includes, “Abraham-Lincoln” (as an example).
As further shown in FIG. 3E, operator node objects (e.g., 373.1) can each refer to another operator node objects (e.g., 372.1) as well as to primitive objects (e.g., Kw3). Thus complex combinations of keyword expression patterns can be defined (built up) with just a small number of operator node objects. The specifying within operator node objects (e.g., 374.1) of primitive patterns can include a specifying of sequence patterns (what comes before or after what; temporally, hierarchically or spatially; and optionally what time gaps or spatial or hierarchical gaps are to be provided there between), a specifying of overlap and/or timing interrelations (what overlaps chronologically or otherwise with what (or does not overlap) and to what extent of overlap or spacing apart) and a specifying of contingent score changing expressions (e.g., IF Kw3 is Near(within 4 words of) Kw4 Then increase matching score or other specified score by indicated amount).
As further shown in FIG. 3E, operator node objects (e.g., 374.1) can uni-directionally or bi-directionally link logically to nodes and/or subregions in other spaces. More specifically, operator node object 374.1 is shown to logically link by way of bi-directional link 370.6 to topic node Tn71 in topic space 313?. Accordingly, if keywords operator node 374.1 is pointed directly to (by matching with it) or pointed to indirectly (by matching to its parent node or child node) by a categorized/normalized CFi or by a plurality of categorized CFi's (e.g., a clustering of CFi's—see 30V.12 of FIG. 3V) or otherwise, then the categorized set of one or more CFi's are thereby logically linked by way of cross-space bi-directional linkages including 370.6 to topic node Tn71. (It is to be noted here that keywords operator node 374.1 does not represent a clustering of CFi's, but rather an operator defined combination of keyword primitives, which combination of primitives may, or may not match to a recently received cluster of CFi's received from a specific user. See also the clustering of CFi's denoted as 30V.12 in FIG. 3V). The cross-space bi-directional link 370.6 in FIG. 3E may have forward direction and/or back direction strength scores associated with it as well as a pointer's-halo size and halo fade factors associated with it so that it (the cross-space link e.g., 370.6) can point to a subregion of the pointed-to other space and not just to a single node within that other space if desired. See also FIGS. 3X and 3Y for enlarged views of how the pointer's-halo size strengths can contribute to total scores of topic nodes (e.g., Tn74? of FIG. 3Y) when a node is painted over by wide projection beams or narrow, focused pointer beams of respective beam intensities (e.g., narrow beam 370.6 sw? in FIG. 3X versus 370.6 sw? in FIG. 3Y). By using a halo'ed pointer. a given operator node can point to and incorporate into itself a collection of adjacent primitives (and/or a collection of adjacent other operator nodes) where the halo'ed pointer may reference a nearby clustering center point (see 371.50 of FIG. 3E), may provide an offset from the clustering center point (see 371.56 of FIG. 3E) and then may specify a radius for a covered circular area centered on that offset point. Other shapes besides encircling circles may be used instead (e.g., ellipses, regular polygons etc.). As used herein, a so-called, pointer's-halo (e.g., the one cast by logical link 370.6? in FIG. 3Y) is not to be confused with a STAN user's ‘touching’ halo although they have a number of similar attributes, such as having variable halo spreads in different hierarchical directions (and/or variable halo spreads in different spatial directions of a multidimensional space that has distance and direction attributes) and such as having variable halo intensities or scoring strengths (positive or negative) and/or variable halo strength fading factors along respective different directions and/or according to respective hierarchical or other radii away from the pointed-to or directly ‘touched’ point in the respective space (e.g., topic space).
While not explicitly shown in FIG. 3E, it is to be understood that operator node objects (e.g., 374.1) can uni-directionally or bi-directionally link logically to informational resources such as chat or other forum participation sessions and/or non-forum research resources and/or to users who cross-associated with the operator node (e.g., an expert or an influencer with regard to the subject matter of the operator node object, e.g., the cognitive sense(s) and the corresponding expression(s) of the operator node). In other words, just as nodes (e.g., Tn71) in topic space can have respective chat rooms cross-associated therewith, operator nodes (e.g., 374.1) in keyword space and/or in other such Cognitive Attention Receiving Spaces can have respective informational resources cross-associated therewith. System users can navigate to a given operator node and can then navigate therefrom to the cross-associated and respective informational resources.
In view of the above, it may be seen that the cross-spaces (inter-space) bi-directional link 370.6 of FIG. 3E may have various strength/intensity attributes logically attached to it for indicating how strongly topic node Tn71 links to operator node object 374.1 and/or how strongly operator node object 374.1 links to topic node Tn71 and/or whether parents (e.g., Tn61) or children (e.g., Tn81) and/or siblings (e.g., Tn74) of the pointed-to topic node Tn71 are also strongly, weakly or not at all linked to the node in the first space (e.g., 370) by virtue of a pointer's-halo cast by link 370.6 (halo not shown in FIG. 3E, see instead FIG. 3X). In other words, by matching or otherwise cross-correlating (e.g., with use of a relative matching or cross-correlating score that does not have to be 100% matching) one or more raw or normalized/categorized CFi's (e.g., clusterings of CFi's) with corresponding nodes in keyword expressions space 370, the STAN—3 system 410 can then automatically discover what nodes (and/or what subregions) of topic space 313? and/or of another space (e.g., context space, emotions space, URL space, etc.) logically link directly or indirectly to the received raw or normalized/categorized CFi's of a given user and how strongly. Linkage scores to different nodes and/or subregions in topic space can be added up for different permutations of CFi's (a.k.a. trial clusterings of CFi's—see 30V.12 of FIG. 3V) and then the topic nodes and/or subregions that score highest can be deemed to be the most likely topic nodes/regions being focused-upon by the STAN user (e.g., user 301A?) from whom the CFi's were collected, and were optionally normalized and/or augmented, clustered into trial permutations and then cross-correlated with similar permutations (e.g., that represented by operator node 374.1) in keyword space. Moreover, linkage scores can be weighted by probability factors where appropriate. Yet more specifically, a first cross-correlation or probability factor may be assigned to a logical linkage (not shown, see 30V.8 of FIG. 3V) as between the keyword combination-and-sequence of node 374.1 and a received clustering of CFi's (e.g., 30V.14 of FIG. 3V) received from a specific user to indicate the likelihood that a received group of keyword expression holding CFi's cross-correlate well with node 374.1. At the same time, a respective other cross-correlation or probability factor may be assigned to another keyword space node to indicate the likelihood that the same received clustering of CFi's (e.g., 30V.14 of FIG. 3V) cross-correlates well with that other node (second keyword space node not shown, but understood to point to a different subregion of topic space than does cross-spaces link 370.6). Then, when corresponding cross-correlation or likelihood scores are automatically computed for competing topic space nodes, the probability factor for each keyword space node is multiplied against the forward pointer strength factor of the corresponding cross-spaces logical link (e.g., that of 370.6) so as to thereby determine the additive (or subtractive) contribution that each cross-spaces logical link (e.g., 370.6) will paint onto the one or more candidate topic nodes it projects its beam (narrow or wide spread beam) on.
The scores contributed by the cross-spaces (inter-space) logical links (e.g., 370.6) need not indicate or merely indicate what candidate topic nodes/subregions the STAN user (e.g., user 301A?) appears to be now focusing-upon based on received raw or categorized CFi's (which received signals can be clustered per FIG. 3V and can point to cross-correlated keyword nodes, i.e. 30V.8 of FIG. 3V; which figure will be detailed later below). They can alternatively or additionally indicate what nodes and/or subregions in user-to-user associations (U2U) space the user (e.g., user 301A?) appears to be focusing-upon and to what degree of likelihood. They can alternatively or additionally indicate what emotions or behavioral states in emotions/behavioral states space the user (e.g., user 301A?) appears to be focusing-upon and to what degree of comparative likelihood. They can alternatively or additionally indicate what context nodes and/or subregions in context space (see 316? of FIG. 3D) the user (e.g., user 301A?) appears to be focusing-upon and to what degree of comparative likelihood. They can alternatively or additionally indicate what context nodes and/or subregions in social dynamics space (see 312? of FIG. 3D) the user (e.g., user 301A?) appears to be focusing-upon and to what degree of comparative likelihood. And so on.
Moreover, linkage strength scores to competing ones of topic nodes (e.g., Tn71 versus Tn74 in the case of FIG. 3E) need not be generated simply on the basis of received CFi's being linked more strongly or weakly to corresponding keyword expression nodes (e.g., 374.1) and the latter being linked more strongly or weakly to one topic node rather than to another (e.g., Tn71 versus Tn74). The cross-spaces linkage strength scores cast from URL nodes in URL space (e.g., the forward strength score going from URL operator node 394.1 to topic node Tn74) can be added in to the accumulating scores of competing ones of topic nodes (e.g., Tn71 versus Tn74). The respective linkage strength scores from Meta-tag nodes in Meta-tag space (395 of FIG. 3E) to the competing topic nodes (e.g., Tn71 versus Tn74) can be included in the machine-implemented computations of competing final scores. The respective linkage strength scores from hybrid nodes (e.g., Kw-Ur node 384.1 linking by way of logical link 380.6) to topic space and/or to another space can be included in the machine-implemented computations of competing final scores. In other words, a rich set of diversified CFi's received from a given STAN user (e.g., user 301A? of FIG. 3D) can be parsed, clustered and cross-correlated to potentially matching (e.g., candidate) points, nodes or subregions in one or more of the system-maintained Cognitive Attention Receiving Spaces and this can lead to a rich set of cross-space linkage scores contributing to (or detracting from) the final scores of different ones of topic nodes so that specific topic nodes and/or topic subregions ultimately become distinguished as being the more likely ones being focused-upon due to the hints and clues collected from the given STAN user (e.g., user 301A? of FIG. 3D) by way of up or in-loaded CFi's, CVi's and the like as well as assistance provided by the then active personal profiles 301 p of the given STAN user (e.g., user 301A? of FIG. 3D).
Cross-spaces logical linkages such as 370.6 (a.k.a. IntEr-Space cross-associating links or “IoS-CAX's”) are referred to herein as “reflective” when they link to a node (e.g., to topic node Tn71) that has additional links back to the same space (e.g., keyword space) from which the first link (e.g., 370.6) came from. Although not shown in FIG. 3E, it is to be understood that a topic node such as Tn71 will typically have more than one logical link (more than just 370.6) logically linking it to nodes in keyword expressions space (as an example) and/or to nodes in other spaces outside of topic space. Accordingly, when a given user's (e.g., user 301A?) CFi's are matched with cross-correlation strength of 100% or less to a first node (e.g., 374.1) in keyword expressions space, that keyword node will likely link to a topic node (e.g., Tn71) that links back to yet other nodes (other than 374.1) in keyword expressions space 370. Therefore, if a cross-correlation is desired as between keyword expressions that have a same topic node or topic space subregion (TSR) in common, the bi-directional nature of cross-spaces links such as 370.6 may be followed to the common nodes in topic space and then a tracing back via other linkages from that region of topic space 313? to keyword expressions space 370 may be carried out by automated machine-implemented means so as to thereby identify the topic-wise cross-correlated other keyword expressions. A similar process may be carried out for identifying URL nodes (e.g., 391.2) that are topic-wise cross-correlated to one another and so on. A similar process may be carried out for identifying URL nodes (e.g., 394.1) that are cross-correlated to each other by way of a common hybrid space node (e.g., 384.1) or by way of a common keyword space node. More generally, cross-correlations as between nodes and/or subregions in one space (e.g., keyword space 370) that have in common, one or more nodes and/or subregions in a second space (e.g., topic space 313? of FIG. 3E) may be automatically discovered by backtracking through the corresponding cross-space linkages (e.g., start at keyword node 374.1, forward track along link 370.6 to topic node Tn71, then chain back to a different node in keyword space 370 by tracking along a different cross-space linkage that logically links node Tn71 to keyword expressions space). In one embodiment, the automated cross-correlations discovering process is configured to unearth the stronger ones of the backlinks from say, common node Tn71 to the space (e.g., 370) where cross-correlations are being sought. One use for this process is to identify better keyword combinations for linking to a given topic space region (TSR) or other space subregion. More specifically, if the Fifth Grade student of the above example had used “Honest Abe” as the keyword combination (see also field 30W.2 of FIG. 3W) for navigating to a topic node directed to the Gettysburg Address (see also data object 30W.14 of FIG. 3W), a search for stronger cross-correlated keyword combinations may inform the student that the keyword combination, “President Abraham Lincoln” would have been a better search expression to be included in the search engine strategy.