US 12,265,786 B2
Systems and methods for automatically serializing and deserializing models
Kenneth Jason Sanchez, San Francisco, CA (US); and Michael Kim, Fairfax, VA (US)
Assigned to QUANATA, LLC, San Francisco, CA (US)
Filed by QUANATA, LLC, San Francisco, CA (US)
Filed on Jun. 3, 2022, as Appl. No. 17/832,133.
Application 17/832,133 is a continuation of application No. 15/784,921, filed on Oct. 16, 2017, granted, now 11,379,655.
Prior Publication US 2022/0292256 A1, Sep. 15, 2022
Int. Cl. G06F 17/00 (2019.01); G06F 40/117 (2020.01); G06F 40/14 (2020.01); G06F 40/186 (2020.01); G06N 20/00 (2019.01); G06Q 40/08 (2012.01)
CPC G06F 40/186 (2020.01) [G06F 40/117 (2020.01); G06F 40/14 (2020.01); G06N 20/00 (2019.01); G06Q 40/08 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A computer system for managing models, the computer system including one or more processors in communication with one or more non-transitory computer-readable media, the computer system configured to:
receive a first model having a plurality of functionalities, wherein the first model comprises a machine learning model;
receive a first template corresponding to the first model, the first template having a plurality of placeholders and a plurality of tags;
identify one or more tags from the plurality of tags by at least comparing the plurality of functionalities in the first model to the plurality of tags in the first template, the one or more tags corresponding to the plurality of functionalities;
generate a human-readable document that describes the first model by at least populating the plurality of placeholders in the first template based upon the one or more tags, the human-readable document describing the plurality of functionalities, the human-readable document being a document separate from the first model, the human-readable document is defined by at least one template, wherein the human-readable document is based on one or more rules or guidelines that are configured to map inputs or outputs of the first model, and wherein to generate the human-readable document comprises to embed a series of hidden identifiers and the one or more tags in the human-readable document;
generate a second model based upon (i) the human-readable document and (ii) the series of hidden identifiers and the one or more tags, as embedded, wherein the second model comprises the plurality of functionalities of the first model; and
output the human-readable document,
wherein the human-readable document includes a plain text description of the first model.