Utilizing passages in fusion-based document retrieval

1. A method implemented in a computer system comprising a processor, memory accessible by the processor, and computer program instructions stored in the memory and executable by the processor, the method comprising:

receiving, at the computer system, a query to retrieve a document from a corpus of documents;

retrieving, at the computer system, a plurality of documents from the corpus of documents using a plurality of retrieval methods, each retrieval method generating a ranked list of retrieved documents and a score for each document;

fusing, at the computer system, the generated ranked list of retrieved documents to form an aggregated ranked list of retrieved documents by:

re-scoring, at the computer system, the plurality of documents according to its passage scores, with respect to the query, using at least one of CombMaxPsg, or PsgPFF, wherein MaxPsg re-scores each document based on a score of the passage with the highest retrieval score, and wherein CombMaxPsg re-scores each document based on a score of the passage with the highest retrieval score combined with a relevance of each passage; and

associating, at the computer system, a given document and its highest scored passage using information indicating a relevance of the passage induced from the plurality of ranked lists.

2. The method of claim 1 , wherein fusing further comprises utilizing, at the computer system, passages contained within each ranked list for estimating a relevance dependent association between pseudo-effective ranked lists, wherein a pseudo-effective ranked list is a ranked list the performance of which is assumed to be effective, and each of the plurality of documents.

3. The method of claim 2 , wherein the retrieving comprises generating an initial fusion score for the retrieved documents using at least one of CombSUM, CombMNZ, Borda, or RRF.

4. The method of claim 3 , wherein the fusing is performed using MaxPsg.

5. The method of claim 4 , wherein MaxPsg re-scores each document based on a score of the passage with the highest retrieval score.

6. A system comprising a processor, memory accessible by the processor, and computer program instructions stored in the memory and executable by the processor to perform:

receiving a query to retrieve a document from a corpus of documents;

retrieving a plurality of documents from the corpus of documents using a plurality of retrieval methods, each retrieval method generating a ranked list of retrieved documents and a score for each document;

fusing the generated ranked list of retrieved documents to form an aggregated ranked list of retrieved documents by:

re-scoring the plurality of documents according to its passage scores, with respect to the query, using at least one of CombMaxPsg, or PsgPFF, wherein MaxPsg re-scores each document based on a score of the passage with the highest retrieval score, and wherein CombMaxPsg re-scores each document based on a score of the passage with the highest retrieval score combined with a relevance of each passage; and

associating a given document and its highest scored passage using information indicating a relevance of the passage induced from the plurality of ranked lists.

7. The system of claim 6 , wherein fusing further comprises utilizing passages contained within each ranked list for estimating a relevance dependent association between pseudo-effective ranked lists, wherein a pseudo-effective ranked list is a ranked list the performance of which is assumed to be effective, and each of the plurality of documents.

8. The system of claim 7 , wherein the retrieving comprises generating an initial fusion score for the retrieved documents using at least one of CombSUM, CombMNZ, Borda, or RRF.

9. The system of claim 8 , wherein the fusing is performed using MaxPsg.

10. The system of claim 9 , wherein MaxPsg re-scores each document based on a score of the passage with the highest retrieval score.

11. A computer program product comprising a non-transitory computer readable storage having program instructions embodied therewith, the program instructions executable by a computer, to cause the computer to perform a method comprising:

receiving, at the computer system, a query to retrieve a document from a corpus of documents;

retrieving, at the computer system, a plurality of documents from the corpus of documents using a plurality of retrieval methods, each retrieval method generating a ranked list of retrieved documents and a score for each document;

fusing, at the computer system, the generated ranked list of retrieved documents to form an aggregated ranked list of retrieved documents by:

re-scoring, at the computer system, the plurality of documents according to its passage scores, with respect to the query, using at least one of CombMaxPsg, or PsgPFF, wherein MaxPsg re-scores each document based on a score of the passage with the highest retrieval score, and wherein CombMaxPsg re-scores each document based on a score of the passage with the highest retrieval score combined with a relevance of each passage; and

associating, at the computer system, a given document and its highest scored passage using information indicating a relevance of the passage induced from the plurality of ranked lists.

12. The method of claim 11 , wherein fusing further comprises utilizing, at the computer system, passages contained within each ranked list for estimating a relevance dependent association between pseudo-effective ranked lists, wherein a pseudo-effective ranked list is a ranked list the performance of which is assumed to be effective, and each of the plurality of documents.

13. The method of claim 12 , wherein the retrieving comprises generating an initial fusion score for the retrieved documents using at least one of CombSUM, CombMNZ, Borda, or RRF.

14. The method of claim 13 , wherein the fusing is performed using MaxPsg.

15. The method of claim 14 , wherein MaxPsg re-scores each document based on a score of the passage with the highest retrieval score.