公开(公告)号：US20230297582A1

公开(公告)日：2023-09-21

申请号：US18325616

申请日：2023-05-30

Applicant: Palantir Technologies Inc.

Inventor： Lawrence Manning ,  Rahul Mehta ,  Daniel Erenrich ,  Guillem Palou Visa ,  Roger Hu ,  Xavier Falco ,  Rowan Gilmore ,  Eli Bingham ,  Jason Prestinario ,  Yifei Huang ,  Daniel Fernandez ,  Jeremy Elser ,  Clayton Sader ,  Rahul Agarwal ,  Matthew Elkherj ,  Nicholas Latourette ,  Aleksandr Zamoshchin

IPC: G06F16/2457 ,  G06F16/35 ,  G06F16/9535 ,  G06F16/28 ,  G06F18/23

CPC classification number: G06F16/24578 ,  G06F16/35 ,  G06F16/9535 ,  G06F16/285 ,  G06F18/23

Abstract: Computer implemented systems and methods are disclosed for automatically clustering and canonically identifying related data in various data structures. Data structures may include a plurality of records, wherein each record is associated with a respective entity. In accordance with some embodiments, the systems and methods further comprise identifying clusters of records associated with a respective entity by grouping the records into pairs, analyzing the respective pairs to determine a probability that both members of the pair relate to a common entity, and identifying a cluster of overlapping pairs to generate a collection of records relating to a common entity. Clusters may further be analyzed to determine canonical names or other properties for the respective entities by analyzing record fields and identifying similarities.

公开(公告)号：US20190213191A1

公开(公告)日：2019-07-11

申请号：US16357108

申请日：2019-03-18

Applicant: Palantir Technologies Inc.

Inventor： James Shuster ,  Daniel Fernandez

IPC: G06F16/2455 ,  G06F8/30 ,  G06F8/38 ,  G06F9/30

CPC classification number: G06F16/24568 ,  G06F8/20 ,  G06F8/313 ,  G06F8/33 ,  G06F8/38 ,  G06F9/30043

Abstract: In an embodiment, a data processing method comprises creating and storing a plurality of analytical notebooks in digital computer storage, wherein each of the analytical notebooks comprises notebook metadata that specifies a kernel for execution, and one or more computational cells, wherein each of the cells comprises cell metadata, a source code reference and an output reference; receiving, in association with a first cell among the one or more cells, first input specifying computer program source code of a function, wherein the function defines an input dataset, a transformation, and one or more variables associated with output data; storing the first cell, excluding the output data, using a first digital data storage system and updating the source code reference to identify the first data storage system; using the kernel specified in the notebook metadata, executing an executable version of the source code to result in generating the output data; storing the output data using a second digital data storage system that is separate from the first digital data storage system and updating the output reference to identify the second data storage system.

公开(公告)号：US20170052958A1

公开(公告)日：2017-02-23

申请号：US15233149

申请日：2016-08-10

Applicant: Palantir Technologies Inc.

Inventor： Lawrence Manning ,  Rahul Mehta ,  Daniel Erenrich ,  Guillem Palou Visa ,  Roger Hu ,  Xavier Falco ,  Rowan Gilmore ,  Eli Bingham ,  Jason Prestinario ,  Yifei Huang ,  Daniel Fernandez ,  Jeremy Elser ,  Clayton Sader ,  Rahul Agarwal ,  Matthew Elkherj ,  Nicholas Latourette ,  Aleksandr Zamoshchin

IPC: G06F17/30

CPC classification number: G06F17/3053 ,  G06F17/30705 ,  G06F17/30867

Abstract: Computer implemented systems and methods are disclosed for automatically clustering and canonically identifying related data in various data structures. Data structures may include a plurality of records, wherein each record is associated with a respective entity. In accordance with some embodiments, the systems and methods further comprise identifying clusters of records associated with a respective entity by grouping the records into pairs, analyzing the respective pairs to determine a probability that both members of the pair relate to a common entity, and identifying a cluster of overlapping pairs to generate a collection of records relating to a common entity. Clusters may further be analyzed to determine canonical names or other properties for the respective entities by analyzing record fields and identifying similarities.

Abstract translation: 公开了计算机实现的系统和方法，用于自动聚类和规范地识别各种数据结构中的相关数据。 数据结构可以包括多个记录，其中每个记录与相应实体相关联。 根据一些实施例，系统和方法还包括通过将记录分组成对来识别与相应实体相关联的记录簇，分析相应的对以确定该对的两个成员与公共实体相关联的概率，以及识别 一组重叠的对以生成与公共实体相关的记录集合。 可以通过分析记录字段并识别相似性来进一步分析集群以确定相应实体的规范名称或其他属性。

公开(公告)号：US11704325B2

公开(公告)日：2023-07-18

申请号：US17812984

申请日：2022-07-15

Applicant: Palantir Technologies Inc.

Inventor： Lawrence Manning ,  Rahul Mehta ,  Daniel Erenrich ,  Guillem Palou Visa ,  Roger Hu ,  Xavier Falco ,  Rowan Gilmore ,  Eli Bingham ,  Jason Prestinario ,  Yifei Huang ,  Daniel Fernandez ,  Jeremy Elser ,  Clayton Sader ,  Rahul Agarwal ,  Matthew Elkherj ,  Nicholas Latourette ,  Aleksandr Zamoshchin

IPC: G06F16/00 ,  G06F16/2457 ,  G06F16/35 ,  G06F16/9535 ,  G06F16/28 ,  G06F18/23

CPC classification number: G06F16/24578 ,  G06F16/285 ,  G06F16/35 ,  G06F16/9535 ,  G06F18/23

Abstract: Computer implemented systems and methods are disclosed for automatically clustering and canonically identifying related data in various data structures. Data structures may include a plurality of records, wherein each record is associated with a respective entity. In accordance with some embodiments, the systems and methods further comprise identifying clusters of records associated with a respective entity by grouping the records into pairs, analyzing the respective pairs to determine a probability that both members of the pair relate to a common entity, and identifying a cluster of overlapping pairs to generate a collection of records relating to a common entity. Clusters may further be analyzed to determine canonical names or other properties for the respective entities by analyzing record fields and identifying similarities.

公开(公告)号：US11593102B1

公开(公告)日：2023-02-28

申请号：US17574505

申请日：2022-01-12

Applicant: Palantir Technologies Inc.

Inventor： James Shuster ,  Daniel Fernandez

IPC: G06F9/44 ,  G06F8/71 ,  G06F40/18 ,  G06F9/455 ,  G06F21/60

Abstract: In an embodiment, a data processing method comprises accessing a computer memory comprising a shareable cell-based computation notebook comprising: notebook metadata specifying a kernel for execution, and a computational cell comprising cell metadata, a source code reference, and an output reference, wherein the cell metadata identifies a particular version of source code of a function that defines an input dataset, a transformation, and one or more variables that are to be associated with output data that is to be generated as a result of executing the particular version of the source code; updating the source code reference to identify a first storage location that is to contain the particular version of the source code of the function; and updating the output reference to identify a second storage location that is to contain the output data that is to be generated as a result of executing the particular version of the source code identified in the cell metadata using the kernel specified in the notebook metadata, wherein the method is performed by one or more computing devices.