Meaning-based advertising and document relevance determination
A method for advertising on electronic data networks based upon semantic differentiation.
1. A method comprising:
receiving a request for one or more advertisements, wherein the request includes a first term;
determining a first location in a semantic space of the first term, wherein the semantic space has a plurality of nodes, each node being associated with a different respective term, and wherein each of the plurality of nodes has a bond with at least one other node in the semantic space, and wherein the bonds between the nodes are assigned respective strengths;
determining a semantic distance between the first location in the semantic space of the first term and a second location in the semantic space of a second term associated with an advertisement, including evaluating bonds between a first node associated with the first term and a second node associated with the second term including evaluating a sum of strengths of the bonds between the first node and the second node;
obtaining an original cost-per-impression amount provided by an advertiser for the second term associated with the advertisement;
modifying the original cost-per-impression amount for the second term provided by the advertiser based on the semantic distance in the semantic space between the first term and the second term, wherein a greater semantic distance results in a lower cost-per-impression amount and wherein a smaller semantic distance results in a higher cost-per-impression amount;
providing the advertisement in response to the request; and
charging the advertiser the modified cost-per-impression amount for the second term instead of the original cost-per-impression amount provided by the advertiser.
2. The method of claim 1 , wherein the modified cost-per-impression amount is inversely proportional to the semantic distance.
3. The method of claim 1 , further comprising:
traversing multiple bonds in the semantic space between the first node associated with the first term and the second node associated with the second term, wherein the semantic distance is based on respective strengths of bonds traversed between the first node and the second node.
4. The method of claim 3 , wherein each bond in the semantic space has a direction of association, and wherein a first strength associated with traversing a bond in a direction opposite to its direction of association is higher than a second strength associated with traversing the bond in its direction of association.
5. The method of claim 1 , wherein modifying the original cost-per-impression amount for the second term provided by the advertiser based on the semantic distance in the semantic space between the first term and the second term comprises modifying the original cost-per-impression amount to be no more than the original cost-per-impression amount.
6. The method of claim 1 , further comprising:
generating a ranking of a plurality of advertisements that are located within a threshold semantic distance to the first location in the semantic space of the first term; and
determining that the advertisement is one of the highest-ranked advertisements in the ranking of the plurality of advertisements.
7. The method of claim 6 , wherein generating the ranking of the plurality of advertisements comprises ordering the plurality of advertisements according to respective semantic distances between the first location of the semantic space of the first term and respective locations of the plurality of advertisements in the semantic space.
8. A system, comprising:
one or more computers; and
a storage device coupled to the one or more computers having instructions stored thereon which, when executed by the one or more computers, cause the one or more computers to perform operations comprising:
receiving a request for one or more advertisements, wherein the request includes a first term;
determining a first location in a semantic space of the first term, wherein the semantic space has a plurality of nodes, each node being associated with a different respective term, and wherein each of the plurality of nodes has a bond with at least one other node in the semantic space, and wherein the bonds between the nodes are assigned respective strengths;
determining a semantic distance between the first location in the semantic space of the first term and a second location in the semantic space of a second term associated with an advertisement, including evaluating bonds between a first node associated with the first term and a second node associated with the second term including evaluating a sum of strengths of the bonds between the first node and the second node;
obtaining an original cost-per-impression amount provided by an advertiser for the second term associated with the advertisement;
modifying the original cost-per-impression amount for the second term provided by the advertiser based on the semantic distance in the semantic space between the first term and the second term, wherein a greater semantic distance results in a lower cost-per-impression amount and wherein a smaller semantic distance results in a higher cost-per-impression amount;
providing the advertisement in response to the request; and
charging the advertiser the modified cost-per-impression amount for the second term instead of the original cost-per-impression amount provided by the advertiser.
9. The system of claim 8 , wherein the modified cost-per-impression amount is inversely proportional to the semantic distance.
10. The system of claim 8 , wherein the operations further comprise:
traversing multiple bonds in the semantic space between the first node associated with the first term and the second node associated with the second term, wherein the semantic distance is based on respective strengths of bonds traversed between the first node and the second node.
11. The system of claim 10 , wherein each bond in the semantic space has a direction of association, and wherein a first strength associated with traversing a bond in a direction opposite to its direction of association is higher than a second strength associated with traversing the bond in its direction of association.
12. The system of claim 8 , wherein modifying the original cost-per-impression amount for the second term provided by the advertiser based on the semantic distance in the semantic space between the first term and the second term comprises modifying the original cost-per-impression amount to be no more than the original cost-per-impression amount.
13. The system of claim 8 , wherein the operations further comprise:
generating a ranking of a plurality of advertisements that are located within a threshold semantic distance to the first location in the semantic space of the first term; and
determining that the advertisement is one of the highest-ranked advertisements in the ranking of the plurality of advertisements.
14. The system of claim 13 , wherein generating the ranking of the plurality of advertisements comprises ordering the plurality of advertisements according to respective semantic distances between the first location of the semantic space of the first term and respective locations of the plurality of advertisements in the semantic space.
15. A computer program product, encoded on one or more non-transitory computer storage media, comprising instructions that when executed by one or more computers cause the one or more computers to perform operations comprising:
receiving a request for one or more advertisements, wherein the request includes a first term;
determining a first location in a semantic space of the first term, wherein the semantic space has a plurality of nodes, each node being associated with a different respective term, and wherein each of the plurality of nodes has a bond with at least one other node in the semantic space, and wherein the bonds between the nodes are assigned respective strengths;
determining a semantic distance between the first location in the semantic space of the first term and a second location in the semantic space of a second term associated with an advertisement, including evaluating bonds between a first node associated with the first term and a second node associated with the second term including evaluating a sum of strengths of the bonds between the first node and the second node;
obtaining an original cost-per-impression amount provided by an advertiser for the second term associated with the advertisement;
modifying the original cost-per-impression amount for the second term provided by the advertiser based on the semantic distance in the semantic space between the first term and the second term, wherein a greater semantic distance results in a lower cost-per-impression amount and wherein a smaller semantic distance results in a higher cost-per-impression amount;
providing the advertisement in response to the request; and
charging the advertiser the modified cost-per-impression amount for the second term instead of the original cost-per-impression amount provided by the advertiser.
16. The computer program product of claim 15 , wherein the modified cost-per-impression amount is inversely proportional to the semantic distance.
17. The computer program product of claim 15 , wherein the operations further comprise:
traversing multiple bonds in the semantic space between the first node associated with the first term and the second node associated with the second term, wherein the semantic distance is based on respective strengths of bonds traversed between the first node and the second node.
18. The computer program product of claim 17 , wherein each bond in the semantic space has a direction of association, and wherein a first strength associated with traversing a bond in a direction opposite to its direction of association is higher than a second strength associated with traversing the bond in its direction of association.
19. The computer program product of claim 15 , wherein modifying the original cost-per-impression amount for the second term provided by the advertiser based on the semantic distance in the semantic space between the first term and the second term comprises modifying the original cost-per-impression amount to be no more than the original cost-per-impression amount.
20. The computer program product of claim 15 , wherein the operations further comprise:
generating a ranking of a plurality of advertisements that are located within a threshold semantic distance to the first location in the semantic space of the first term; and
determining that the advertisement is one of the highest-ranked advertisements in the ranking of the plurality of advertisements.
21. The computer program product of claim 20 , wherein generating the ranking of the plurality of advertisements comprises ordering the plurality of advertisements according to respective semantic distances between the first location of the semantic space of the first term and respective locations of the plurality of advertisements in the semantic space.