Multiple rounds of results summarization for improved latency and relevance

Patent metadata

Field	Value
Publication number	US-10095783-B2
Application number	US-201514720964-A
Country	US
Kind code	B2
Filing date	May 25, 2015
Priority date	May 25, 2015
Publication date	Oct 9, 2018
Grant date	Oct 9, 2018

Abstract

Architecture that splits the generation of results summaries into at least two phases for reduced latency and improved relevance. A first phase generates a summary quickly and thereby enables subsequent modules to begin processing. A second round then executes in parallel to these other modules to offset the latency. The second round can also provide additional contextual information to the summarization module to improve the quality (relevance) of the summaries. Ultimately, the summaries generated in the two phases are merged. The multiple rounds of summarization enable the first round to be cheaper in processing resources to save overall latency, whereas the second phase can be more expensive, since the second phase can be run in parallel with other modules. Additionally, contextual information can be used to build the full content summaries in the second round.

First claim

What is claimed is:

1. A system, comprising:
a summarization phase component configured to implement multiple summarization phases of search results determined to be relevant to a submitted query, the multiple summarization phases comprise a first summarization phase and a second summarization phase, the second summarization phase commenced relative to initiation of the first summarization phase;
a summarization component configured to process the multiple summarization phases of the results to generate preliminary results summaries for the first summarization phase and second results summaries for the second summarization phase;
a merge component to merge the preliminary results summaries for the first summarization phase and the second results summaries for the second summarization phase to create merged summaries prior to returning search results to a user device; and
at least one hardware processor configured to execute computer-executable instructions in a memory, the instructions executed to enable the summarization phase component and the summarization component.

2. The system of claim 1 , further comprising a selection component configured to select a top result summary from the merged summaries.

3. The system of claim 1 , wherein the summarization component receives additional contextual information extracted from the preliminary summarization results to generate the second results summaries.

4. The system of claim 1 , wherein the summarization component is configured to identify the first summarization phase and process the first summarization phase in a way different than processing of the second summarization phase.

5. The system of claim 4 , wherein the summarization component employs a first set of algorithms for processing of the first summarization phase and a different set of algorithms for processing of the second summarization phase.

6. The system of claim 1 , wherein the second summarization phase is commenced to overlap the first summarization phase.

7. The system of claim 1 , wherein the second summarization phase is commenced after completion of the first summarization phase, and receives summaries from the first summarization phase for processing in the second summarization phase.

8. A method for reducing latency in creating search results or improving the relevance of search results or both, comprising acts of:
initiating a preliminary summarization phase and initiating a second summarization phase either prior to receiving preliminary results summaries from the preliminary summarization phase or in response to receiving preliminary results summaries from the preliminary summarization phase;
receiving the preliminary results summaries from the preliminary summarization phase without sending the preliminary results summaries to the client device, the preliminary results summaries comprising summaries of search results from a user query;
receiving second results summaries from the second summarization phase, the second results summaries comprising summaries of search results from the user query; and
prior to returning results to the client device, merging the preliminary results summaries and the second results summaries into merged summaries for final summarization processing.

9. The method of claim 8 , further comprising presenting a top set of summaries from the merged summaries, in a results page to the client device.

10. The method of claim 8 , further comprising generating the second results summaries using the preliminary results summaries as input to the second summarization phase.

11. The method of claim 8 , further comprising employing a first set of algorithms to generate the preliminary results summaries from the preliminary summarization phase, and a different set of algorithms to generate full content second results summaries from the second summarization phase.

12. The method of claim 8 , further comprising processing a request that identifies the preliminary summarization phase or the second summarization phase.

13. The method of claim 12 , further comprising automatically switching between the preliminary summarization phase and the second summarization phase based on the request.

14. The method of claim 8 , wherein merging the preliminary results summaries and second results summaries comprises at least one of:
replacing the preliminary results summaries with the second results summaries;
combining the preliminary results summaries with the second results summaries; and
deciding to keep one of the preliminary results summaries and the second results summaries and discarding the other.

15. A method that reduces latency in generating search results, comprising acts of:
generating multiple summarization phases of search results, comprising a first summarization phase and a second summarization phase, the second summarization phase commenced relative to initiation of the first summarization phase;
processing the multiple summarization phases of the search results to generate preliminary results summaries for the first summarization phase and second results summaries for the second summarization phase; and
prior to returning results to a client device, merging the preliminary results summaries and the second results summaries into merged summaries for final summarization processing.

16. The method of claim 15 , further comprising processing the second summarization phase while the first summarization phase is being processed or after the first summarization phase completes.

17. The method of claim 15 , further comprising generating the preliminary results summaries with less content than the second results summaries to reduce latency in succeeding dependent processing of the preliminary results summaries.

18. The method of claim 15 , further comprising generating the second results summaries using the preliminary results summaries as input to the second summarization phase.

19. The method of claim 15 , further comprising presenting a top set of results summaries from the merged summaries in a results page.

20. The method of claim 17 , wherein the preliminary results are sent for additional dependent processing that is dependent on the reduced content preliminary results being available.

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Donato Gianluca
• Avanesyan Ara
• Rajaraman Aparna

Classifications

• G06F16/3329

  Natural language query formulation · CPC title

• G06F16/9577 Primary

  Optimising the visualization of content, e.g. distillation of HTML documents · CPC title

• G06F40/174

  Form filling; Merging · CPC title

• G06F16/345 Primary

  Summarisation for human users · CPC title

• G06F17/30719 Primary

  Physics · mapped topic