Context-based search visualization and context management using neural networks

• US 7,475,072 B1
• Filed: 08/30/2006
• Issued: 01/06/2009
• Est. Priority Date: 09/26/2005
• Status: Active Grant

• Current Assignee: Callahan Cellular LLC (Intellectual Ventures LLC)
• Original Assignee: Quintura Inc.
• Inventors: Ershov, Alexander V.
• Primary Examiner(s): Mahmoudi; Tony
• Assistant Examiner(s): McCue; Brittany N
• Application Number:
  US11/468,692
• Time in Patent Office: 860 Days
• Field of Search: 707/6
• US Class Current: 1/1
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

A system, method and computer program product for visualization of context-based search results, including a plurality of neurons, the neurons being associated with words and documents; a plurality of connections between the neurons; a map that displays at least some of the neurons to a user, wherein the display of the neurons on the map corresponds to their relevance to a search query; a display of the links to the relevant documents; and means for changing positions of the neurons relative to each other based on input from the user. Changing a position of one neuron relative to other neurons also changes positions of other contextually relevant neurons, and displays different relevant documents. The map displays the neurons with their relevance identified by any of font type, color, transparency and font size. The map includes icons in proximity to the displayed word neurons for identifying those neurons as irrelevant. Links to the documents are obtained from a search engine having an input query. The map displays annotations and/or keywords to the documents next to the displayed documents.

Claims

1. A system for visualization of context-based search results comprising:

a processor;

a memory operatively coupled to the processor;

computer code in the memory for implementing the following functionality;

a) a plurality of neurons that form a neural network, the neurons being associated with documents and words in the documents, the neurons representing words corresponding to a search query (“

word-neurons”

), and the neurons representing documents corresponding to the search results (“

document-neurons”

);

b) a plurality of connections between the word-neurons and the document-neurons;

c) means for displaying the word-neurons and the search results to a user,

wherein relevance of the word-neurons is indicated by any of font size, text transparency, font color, and animation effects associated with text of at least some of the neurons; and

d) means for changing positions of the word-neurons on a display relative to each other based on input from the user,

wherein changing a position of one word-neuron relative to other word-neurons also changes positions of other contextually relevant word-neurons, thereby changing a context of the search query and permitting the user to identify different relevant documents, and

wherein at least some of the neurons that correspond to objects (concepts) have icons associated with them, the icons corresponding to a meaning of the objects.

2. The system of claim 1, wherein the word-neurons form a first layer, wherein the document-neurons form a second layer, and further comprising a layer of object (concept) neurons and a layer of sentence neurons, and wherein each neuron is connected to some, but not all, of the other neurons.

3. The system of claim 1, wherein the input from the user comprises dragging a word-neuron on a display device.

4. The system of claim 1, wherein the input from the user comprises changing a weight coefficient of a connection between two word-neurons to indicate relevance.

5. The system of claim 1, wherein the word-neurons are displayed in a form of any of:

(a) a two-dimensional “

mountain peak”

diagram, with the most relevant word-neurons located at the top of the diagram, (b) positioned within a pyramid, with the most relevant word-neurons located at the base of the pyramid, and the most irrelevant word-neurons located at the point of the pyramid, (c) positioned within a sphere, with the most relevant word-neurons located on a surface of the sphere, and the most irrelevant word-neurons located at the center of the sphere, (d) positioned within a sphere, with the most irrelevant word-neurons located on a surface of the sphere, and the most relevant word-neurons located at the center of the sphere, (e) positioned within a cube, with the most relevant word-neurons located on a surface of the cube, and the most irrelevant word-neurons located at the center of the cube, (f) positioned within a polyhedron, with the most relevant word-neurons located on a surface of the polyhedron, and the most irrelevant word-neurons located at the center of the polyhedron, (g) positioned within a polyhedron, with the most irrelevant word-neurons located on a surface of the polyhedron, and the most relevant word-neurons located at the center of the polyhedron.

6. A system for visualization of context-based search results comprising:

a processor;

a memory operatively coupled to the processor;

computer code in the memory for implementing the following functionality;

a) a plurality of neurons that form a neural network, the neurons being associated with words and documents, the neurons representing words corresponding to a search query (“

word-neurons”

), and the neurons representing documents corresponding to the search results (“

document-neurons”

);

b) a plurality of connections between the word-neurons and the document-neurons, wherein only some of the word-neurons are connected to each other;

c) means for displaying the word-neurons to a user, wherein relevance of the word-neurons is indicated by any of font size, text transparency, font color, and animation effects associated with text of at least some of the neurons; and

d) means for changing positions of the word-neurons relative to each other based on input from the user,

wherein changing a position of one word-neuron relative to other word-neurons also changes positions of other contextually relevant word-neurons, thereby changing a context of the search query and permitting the user to identify different relevant documents, and

wherein at least some of the neurons that correspond to objects (concepts) have icons associated with them, the icons corresponding to a meaning of the objects.

7. A method for visualization of context-based search results comprising:

a) displaying a plurality of neurons that form a neural network, the neurons being associated with words, objects, sentences and documents, the neurons representing words corresponding to a search query (“

word-neurons”

), the neurons representing objects corresponding to concepts (“

object-neurons”

), the neurons representing sentences corresponding to sentences in the documents (“

sentence-neurons”

) and the neurons representing documents corresponding to the search results (“

document-neurons”

),

wherein relevance of the word-neurons is indicated by any of font size, text transparency, font color, and animation effects associated with text of at least some of the neurons, and

wherein at least some of the neurons that correspond to objects (concepts) have icons associated with them, the icons corresponding to a meaning of the objects;

b) based on user input, changing a position of one word-neuron; and

c) in response to the change in position of one word-neuron, changing positions of other word-neurons on a display, thereby changing a context of the search query so as to identify contextually relevant documents.

8. The method of claim 7, wherein step (b) comprises pulling a relevant word-neuron upwards.

9. The method of claim 7, wherein step (b) comprises pushing an irrelevant neuron downwards.

10. The method of claim 7, wherein step (b) comprises changing a weight of a coefficient of a connection between two word-neurons.

11. The method of claim 7, wherein step (a) comprises displaying the word-neurons as being positioned on any of a circle, a sphere, a pyramid, a cube and a polyhedron.

12. A system for visualization of context-based search results comprising:

a processor;

a memory operatively coupled to the processor;

computer code in the memory for implementing the following functionality;

a) a plurality of neurons forming a neural network, the neurons being associated with words and documents, the neurons representing words corresponding to a search query (“

word-neurons”

), and the neurons representing documents corresponding to the search results (“

document-neurons”

);

b) a plurality of connections between the word-neurons and the document-neurons;

c) a map that displays at least some of the word-neurons and the document-neurons to a user, wherein a manner of display and locations of the word-neurons on the map corresponds to their contextual relevance to a search query,

wherein relevance of the word-neurons is indicated by any of font size, text transparency, font color, and animation effects associated with text of at least some of the neurons, and

wherein at least some of the neurons that correspond to objects (concepts) have icons associated with them, the icons corresponding to a meaning of the objects;

d) a display of hyperlinks to the relevant documents; and

e) means for changing positions of the word-neurons relative to each other based on input from the user,

f) wherein changing a position of one word-neuron relative to other word-neurons also changes positions of other contextually relevant word-neurons, thereby changing a context of the search query and displays different relevant documents.

13. The system of claim 12, wherein the map displays the word-neurons with their relevance identified by any of font type, color, transparency and font size.

14. The system of claim 12, wherein the map includes icons in proximity to the displayed word-neurons for identifying those word-neurons as irrelevant.

15. The system of claim 12, wherein the hyperlinks to the documents are obtained from a search engine having an input query.

16. The system of claim 12, wherein the map displays annotations for the documents next to the displayed documents.

17. The system of claim 12, wherein if a cursor is held over a word neuron on the map, the map displays other word-neurons that have connections to that word, and which were not previously displayed.

18. The system of claim 17, wherein, once the cursor is removed from the word neuron, the other word-neurons are no longer displayed.

19. The system of claim 17, wherein different search results are displayed based on a newly formed search query that includes the word-neuron, while the cursor is being held over the word-neuron.

20. The system of claim 17, wherein, once the cursor is removed from the word-neuron, the other word-neurons are no longer displayed, and previously displayed search results are again displayed.

21. A system for managing context of a search comprising:

a processor;

a memory operatively coupled to the processor;

computer code in the memory for implementing the following functionality;

a) a map that graphically displays search terms to a user in response to a search query, wherein the search terms are each associated with a word, and relative positions of the words corresponds to their connections in documents identified as relevant to the search query,

wherein relevance of the word-neurons is indicated by any of font size, text transparency, font color, and animation effects associated with text of at least some of the neurons, and

wherein at least some of the neurons that correspond to objects (concepts) have icons associated with them, the icons corresponding to a meaning of the objects; and

b) a display of search results corresponding to the search query;

c) wherein, bringing a cursor over one of the words temporarily displays other words, not previously shown, that are contextually connected to that word thereby changing a context of the search query and changes the display of the documents identified as most relevant, while the cursor is held over the selected word, and

d) wherein selecting the word permanently adds the selected word to the search query thereby changing a context of the search query and correspondingly changes the display of the documents identified as most relevant.

22. The system of claim 21, wherein the display of the search results includes a display of hyperlinks associated with the documents and a display of annotations associated with each document shown in proximity to the each document.

23. The system of claim 21, wherein the display of the search results includes a display of hyperlinks associated with the documents and a display of keywords associated with each document shown in proximity to the each document.

24. The system of claim 21, wherein the words displayed on the map at any given time to the user are associated with a corresponding word-neuron, such that connections between the word-neurons are reflected in the relative positions of the word-neurons on the map.

25. The system of claim 21, wherein a context of the displayed search results is saved for later retrieval.

26. A method for managing context of a search comprising:

a) graphically displaying on a map, to a user, search terms in response to a search query, wherein each search term is associated with a word, and relative positions of the words on the map corresponds to their connections in documents identified as relevant to the search query,

wherein relevance of the words is indicated by any of font size, text transparency, font color, and animation effects associated with text of at least some of the words, and

wherein at least some of the neurons that correspond to objects (concepts) have icons associated with them, the icons corresponding to a meaning of the objects; and

b) displaying search results corresponding to the search query;

c) wherein, bringing a cursor over one of the words temporarily displays other words, not previously shown, that are contextually connected to that word, thereby changing a context of the search query, and changes the display of the documents identified as most relevant, while the cursor is held over the selected word, and

d) wherein selecting the word permanently adds the selected word to the search query thereby changing a context of the search query and correspondingly changes the display of the documents identified as most relevant.

27. A method for managing context of a search comprising:

a) receiving a search query from a user;

b) generating search results based on the search query, the search results corresponding to identified documents;

c) associating a plurality of neurons that form a neural network with the documents and words in the documents, the neurons representing words corresponding to a search query (“

word-neurons”

), and the neurons representing documents corresponding to the search results (“

document-neurons”

);

d) forming a plurality of connections between the word-neurons and the document-neurons;

e) displaying word-neurons corresponding to the search query on a map, wherein relative positions of the word-neurons, on the map, correspond to their connections in documents identified as relevant to the search query,

wherein relevance of the word-neurons is indicated by any of font size, text transparency, font color, and animation effects associated with text of at least some of the neurons, and

wherein at least some of the neurons that correspond to objects (concepts) have icons associated with them, the icons corresponding to a meaning of the objects;

f) changing a position of one word-neuron on the map relative to other word-neurons to thereby identify more relevant documents of the identified documents; and

g) displaying, to the user, the documents considered most relevant based on the search query and the relative positions of the displayed word-neurons on the map.

28. The method of claim 27, wherein bringing a cursor over one of the word-neurons on the map temporarily displays other words, not previously shown, that are contextually connected to that word-neuron, thereby changing a context of the search query and changes the display of the documents identified as most relevant, while the cursor is held over the selected word-neuron, and

wherein selecting the word-neuron permanently adds the selected word-neuron to the search query, thereby changing a context of the search query, and correspondingly changes the display of the documents identified as most relevant.

29. A system for managing context of a search comprising:

a processor;

a memory operatively coupled to the processor;

computer code in the memory for implementing the following functionality;

a) means for receiving a search query from a user;

b) means for generating search results based on the search query, the search results corresponding to identified documents;

c) means for associating a plurality of neurons that form a neural network with the documents and words in the documents, the neurons representing words corresponding to the search query (“

word-neurons”

), and the neurons representing documents corresponding to the search results (“

document-neurons”

);

d) means for forming a plurality of connections between the word-neurons and the document-neurons;

e) means for displaying the word-neurons corresponding to the search query on a map, wherein relative positions of the words, on the map, correspond to their connections in documents identified as relevant to the search query,

wherein relevance of the word-neurons is indicated by any of font size, text transparency, font color, and animation effects associated with text of at least some of the neurons, and

wherein at least some of the neurons that correspond to objects (concepts) have icons associated with them, the icons corresponding to a meaning of the objects;

f) means for changing a position of one word-neuron on the map relative to other word-neurons to thereby change a context of the search query and identify more relevant documents of the identified documents; and

g) means for displaying, to the user, the documents considered most relevant based on the search query and based on the relative positions of the displayed word-neurons on the map.

30. A system for visualization of context-based search results comprising:

a processor;

a memory operatively coupled to the processor;

computer code in the memory for implementing the following functionality;

a) a plurality of neurons forming a neural network, the neurons being associated with documents, words in the documents and sentences in the documents, the neurons representing words corresponding to a search query (“

word-neurons”

), the neurons representing sentences corresponding to sentences in the documents (“

sentence-neurons”

) and the neurons representing documents corresponding to the search results (“

document-neurons”

), wherein the documents are identified in response to a search query from a user;

b) a plurality of connections between the word-neurons and the document-neurons, wherein there are connections between some, but not all, word-neurons;

c) a map displaying the word-neurons to a user, documents identified as most relevant and annotations for the documents identified as most relevant,

wherein relevance of the word-neurons is indicated by any of font size, text transparency, font color, and animation effects associated with text of at least some of the neurons; and

d) means for changing positions of the word-neurons relative to each other based on input from the user,

wherein changing a position of one word-neuron relative to other neurons also changes positions of other contextually relevant word-neurons, thereby changing a context of the search query and permitting the user to identify other documents that are more relevant to the user'"'"'s search, and

wherein at least some of the neurons that correspond to objects (concepts) have icons associated with them, the icons corresponding to a meaning of the objects.