IP Library Granted Patent US 12,517,913
Granted Patent B1
US 12,517,913 · App. 18/986,521 · Granted Jan 6, 2026

Performing unified search using a hybrid search index

Inventors: Devin Mancuso (San Francisco, CA); Alan Chu (Arvada, CO); Brett Bergeron (Portland, OR); Maor Bar Asher (New York, NY); Ryan YeePin Yheng (Milpitas, CA); Shweta Kode (Sammamish, WA)
Assignee: Dropbox, Inc.
G06F16/24578G06F16/248
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,517,913
App. No.
18/986,521
Granted
Jan 6, 2026
Kind
B1
Abstract

The present disclosure is directed toward systems, methods, and non-transitory computer readable media for generating and searching a hybrid search index. In some embodiments, the disclosed systems generate a hybrid search index that comprises one or more content items stored at a content management system or at external network locations linked to the content management system via software connectors along with world state data associated with the one or more content items. The disclosed systems can generate a search result from the hybrid search index in response to receiving a search query of the hybrid search index. In some cases, the disclosed systems can rank one or more content items included in the search result based on observation layer data of the one or more content items.

Claims (54)

1 . A computer-implemented method comprising:

receiving, from a client device, a search query for searching a hybrid search index;

generating a search result for the search query by utilizing the hybrid search index to access a first content item from a first storage location on a first network and a second content item from a second storage location on a second network;

generating, for display on the client device, a search result interface comprising the search result, a first rationale describing reasoning for including the first content item in the search result, and a second rationale describing reasoning for including the second content item in the search result, wherein the search result interface comprises a first summarization element selectable to generate a first content item summary; and

generating, in response to a selection of the first summarization element, the first content item summary by utilizing a large language model to determine and summarize portions of the first content item relevant to a user account associated with the client device.

2 . The computer-implemented method of claim 1 , further comprising generating the first rationale by utilizing a large language model to generate text describing the reasoning for including the first content item in the search result based on one or more of world state data, observational layer data, or user interaction data stored within the hybrid search index.

3 . The computer-implemented method of claim 1 ,

further comprising:

determining an event associated with the user account associated with the client device and related to the search query; and

generating the first rationale describing reasoning for including the first content item in the search result, wherein the first rationale comprises information relating to the event associated with the user account.

4 . The computer-implemented method of claim 1 , further comprising:

determining a relevancy ranking of the first content item in the search result; and

automatically generating the first content item summary for the first content item based on the relevancy ranking of the first content item.

5 . The computer-implemented method of claim 1 , further comprising providing, for display within the search result interface, a search result summarization element selectable to summarize the search result.

6 . The computer-implemented method of claim 5 , further comprising generating, in response to a selection of the search result summarization element, a search result summary by utilizing a large language model to summarize content from a subset of content items included within the search result.

7 . The computer-implemented method of claim 1 , further comprising:

providing, for display within the search result interface, a search result input field selectable to enter search result questions regarding one or more content items included in the search result; and

generating, in response to a search result question entered via the search result input field, a search result response by utilizing a large language model to generate text responsive to the search result question based on analyzing at least one content item referenced by the search result question.

8 . A system comprising:

at least one processor; and

a non-transitory computer readable medium comprising instructions that,

when executed by the at least one processor, cause the system to:

receive, from a client device, a search query for searching a hybrid search index;

generate a search result for the search query by utilizing the hybrid search index to access a first content item from a first storage location on a first network and a second content item from a second storage location on a second network;

generate, for display on the client device, a search result interface comprising the search result, a first summarization element selectable to summarize the first content item in the search result, and a second summarization element selectable to summarize the second content item in the search result;

generate, in response to a selection of the first summarization element, a first content item summary by utilizing a large language model to summarize content from the first content item; and

generate the first content item summary by utilizing the large language model to determine and summarize portions of the first content item relevant to a user account associated with the client device.

9 . The system of claim 8 , further comprising instructions that, when executed by the at least one processor, cause the system to:

determine an event associated with the user account associated with the client device and related to the search query; and

generate a first rationale describing reasoning for including the first content item in the search result, wherein the first rationale comprises information relating to the event associated with the user account.

10 . The system of claim 8 , further comprising instructions that, when executed by the at least one processor, cause the system to:

determine a relevancy ranking of the first content item in the search result; and

automatically generate the first content item summary for the first content item based on the relevancy ranking of the first content item.

11 . The system of claim 8 , further comprising instructions that, when executed by the at least one processor, cause the system to generate, for display within search result interface, a first rationale describing reasoning for including the first content item in the search result, and a second rationale describing reasoning for including the second content item in the search result.

12 . The system of claim 11 , further comprising instructions that, when executed by the at least one processor, cause the system to generate the first rationale for the first content item by utilizing a large language model to generate text describing the reasoning for including the first content item based on one or more of world state data, observational layer data, or user interaction data stored within the hybrid search index.

13 . The system of claim 8 , further comprising instructions that, when executed by the at least one processor, cause the system to generate the search result by accessing the first content item using a software connector linking an external application hosting the first content item to a content management system housing the hybrid search index.

14 . The system of claim 8 , further comprising instructions that, when executed by the at least one processor, cause the system to:

provide, for display within the search result interface, a search result input field selectable to enter search result questions regarding one or more content items included in the search result; and

generate, in response to a search result question entered via the search result input field, a search result response by utilizing a large language model to generate text responsive to the search result question based on analyzing at least one content item referenced by the search result question.

15 . A non-transitory computer readable medium comprising instructions that, when executed by at least one processor, cause the at least one processor to:

receive, from a client device, a search query for searching a hybrid search index;

generate a search result for the search query by utilizing the hybrid search index to access a first content item from a first storage location on a first network and a second content item from a second storage location on a second network utilizing software connectors linking external applications to a content management system housing the hybrid search index;

generate, for display on the client device, a search result interface comprising the search result, a first rationale describing reasoning for including the first content item in the search result, and a second rationale describing reasoning for including the second content item in the search result;

generate, for display within the search result interface, a first summarization element selectable to generate a first content item summary for the first content item and a second summarization element selectable to generate a second content item summary for the second content item; and

generate, in response to a selection of the first summarization element, the first content item summary by utilizing a large language model to determine and summarize portions of the first content item relevant to a user account associated with the client device.

16 . The non-transitory computer readable medium of claim 15 , further comprising instructions that, when executed by the at least one processor, cause the at least one processor to generate the first rationale by utilizing a large language model to generate text describing the reasoning for including the first content item in the search result based on one or more of world state data, observational layer data, or user interaction data stored within the hybrid search index.

17 . The non-transitory computer readable medium of claim 15 , further comprising instructions that, when executed by the at least one processor, cause the at least one processor to:

determine an event associated with the user account associated with the client device and related to the search query; and

generate the first rationale describing reasoning for including the first content item in the search result, wherein the first rationale comprises information relating to the event associated with the user account.

18 . The non-transitory computer readable medium of claim 15 , further comprising instructions that, when executed by the at least one processor, cause the at least one processor to:

determine a relevancy ranking of the first content item in the search result; and

automatically generate the first content item summary for the first content item based on the relevancy ranking of the first content item.

19 . The non-transitory computer readable medium of claim 15 , further comprising instructions that, when executed by the at least one processor, cause the at least one processor to provide, for display within the search result interface, a search result summarization element selectable to summarize the search result.

20 . The non-transitory computer readable medium of claim 19 , further comprising instructions that, when executed by the at least one processor, cause the at least one processor to generate, in response to a selection of the search result summarization element, a search result summary by utilizing a large language model to summarize content from a subset of content items included within the search result.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jan 2, 2025
From: MANCUSO, DEVIN; CHU, ALAN; BERGERON, BRETT; ASHER, MAOR BAR; YHENG, RYAN YEEPIN; KODE, SHWETA
To: DROPBOX, INC.
Reel/Frame 069723/0879 →
Continuity (1)
Continuation 18815619 · Aug 26, 2024
References Cited (49)
US 8190541B2 · Downs et al. · 2012 [cited by applicant]
US 8688669B1 · Bernstein · 2014 [cited by applicant]
US 9037579B2 · Pasumarthi et al. · 2015 [cited by applicant]
US 9043358B2 · Cheng et al. · 2015 [cited by applicant]
US 9201889B1 · Madany et al. · 2015 [cited by applicant]
US 9547696B2 · Seth et al. · 2017 [cited by applicant]
US 9558248B2 · Kirazci et al. · 2017 [cited by applicant]
US 9805097B2 · Chang et al. · 2017 [cited by applicant]
US 10318570B2 · Billi-Duran et al. · 2019 [cited by applicant]
US 10528021B2 · Billi et al. · 2020 [cited by applicant]
US 10545492B2 · Billi-Duran et al. · 2020 [cited by applicant]
US 10713307B2 · Bliss et al. · 2020 [cited by applicant]
US 10795895B1 · Taig et al. · 2020 [cited by applicant]
US 11531712B2 · Gupta et al. · 2022 [cited by applicant]
US 11537671B1 · Gray et al. · 2022 [cited by applicant]
US 11645671B2 · Raviv et al. · 2023 [cited by applicant]
US 11720920B1 · Zhou et al. · 2023 [cited by applicant]
US 11741115B2 · Gutiérrez et al. · 2023 [cited by applicant]
US 20120077521A1 · Boldyrev et al. · 2012 [cited by applicant]
US 20140379661A1 · Marcus et al. · 2014 [cited by applicant]
US 20150248410A1 · Stickler et al. · 2015 [cited by applicant]
US 20160132538A1 · Bliss et al. · 2016 [cited by applicant]
US 20160196360A1 · Best et al. · 2016 [cited by applicant]
US 20160224682A1 · Iyer et al. · 2016 [cited by applicant]
US 20160373891A1 · Ramer et al. · 2016 [cited by applicant]
US 20170075997A1 · Wan et al. · 2017 [cited by applicant]
US 20170097933A1 · Rajan et al. · 2017 [cited by applicant]
US 20170351226A1 · Bliss et al. · 2017 [cited by applicant]
US 20170351703A1 · Shreyas et al. · 2017 [cited by applicant]
US 20180075034A1 · Wang et al. · 2018 [cited by applicant]
US 20200334305A1 · Kussmaul et al. · 2020 [cited by applicant]
US 20200334313A1 · Kussmaul et al. · 2020 [cited by applicant]
US 20210150100A1 · Laloi · 2021 [cited by applicant]
US 20210248643A1 · Zhang et al. · 2021 [cited by applicant]
US 20210406332A1 · Helvik et al. · 2021 [cited by applicant]
US 20220156262A1 · Chen et al. · 2022 [cited by applicant]
US 20220414170A1 · Helvik et al. · 2022 [cited by applicant]
US 20230409581A1 · Betthauser et al. · 2023 [cited by applicant]
US 20240202539A1 · Poirier et al. · 2024 [cited by applicant]
US 20240256874A1 · Yousefi Maragheh et al. · 2024 [cited by applicant]
US 20240273145A1 · Baek et al. · 2024 [cited by applicant]
US 20240281487A1 · Bathwal et al. · 2024 [cited by applicant]
CN 110941641B · 2022 [cited by applicant]
CN 111382226B · 2023 [cited by applicant]
Rahimi et al., “Explaining Documents' Relevance to Search Queries”, Nov. 2, 2021, J. ACM, vol. 0, No. 0, pp. 1-24 (Year: 2021). [cited by examiner]
“Exalead CloudViewTM Platform Highlights,” Dassault Systemes, 3DS.com/Exalead, 2012, 20 pages, Retrieved from the Internet: URL: https://www.3ds.com/fileadmin/Industnes/Federal/3ds-exalead_product-sheet_cloudview-platfo… [cited by applicant]
Non-Final Office Action from U.S. Appl. No. 18/815,606, mailed Nov. 20, 2024, 18 pages. [cited by applicant]
Notice of Allowance for U.S. Appl. No. 18/815,619 mailed on Nov. 15, 2024, 24 pages. [cited by applicant]
Notice of Allowance for U.S. Appl. No. 18/815,619 mailed on Jan. 16, 2025, 7 pages. [cited by applicant]
Cited By (1)
US 12,670,160