US 12,229,665 B2
Secure search engine utilizing a learning engine
Gregory Austin Allison, Napa, CA (US); Douglas Anderson, San Francisco, CA (US); and Daivak Sunil Shah, Fremont, CA (US)
Assigned to Pacaso Inc., Cincinnati, OH (US)
Filed by Pacaso Inc., Cincinnati, OH (US)
Filed on Jan. 29, 2021, as Appl. No. 17/163,192.
Prior Publication US 2022/0245442 A1, Aug. 4, 2022
Int. Cl. G06N 3/08 (2023.01); G06F 16/2457 (2019.01); G06F 16/2458 (2019.01); G06F 16/29 (2019.01); G06N 3/044 (2023.01); G06N 3/063 (2023.01)
CPC G06N 3/08 (2013.01) [G06F 16/24578 (2019.01); G06F 16/2477 (2019.01); G06F 16/29 (2019.01); G06N 3/044 (2023.01); G06N 3/063 (2013.01)] 30 Claims
OG exemplary drawing
 
1. A system configured to enable searches for physical resources, comprising:
a network interface;
an application specific integrated circuit (ASIC), implementing a first neural network, the ASIC comprising:
one or more long short-term memory units configured to keep track of dependencies between an elements in an input sequence, the one or more long short-term memory units comprising:
a cell, and
regulators, the regulators including an input gate, an output gate and a forget gate;
at least one processing device operable to:
access attributes of a plurality of physical resources,
wherein a given physical resource in the plurality of physical resources is configured to be shared amongst a plurality of physical resource users in a time displaced manner;
receive over a network, using the network interface, a communication providing attributes for a first user;
search for physical resources in the plurality of physical resources that correspond with attributes of the first user, wherein a second neural network is used to identity a first set of physical resources in the plurality of physical resources using the attributes of the first user and the attributes of the plurality of physical resources;
generate and associate respective match scores with physical resources in the first set of physical resources identified using the second neural network;
identify a subset of physical resources comprising physical resources in the first set of physical resources that at least one other user of the system has received a right to use;
identify, using the first neural network, users associated with the subset of physical resources that have at least a first likelihood of temporal usage conflicts with the first user;
based at least in part on the identified users associated with the subset of physical resources that have at least the first likelihood of temporal usage conflicts with the first user, adjusting one or more match scores associated with one or more physical resources in the subset of physical resources;
generate, based at least in part on the adjusted match scores, ranked search results for the first set of physical resources identified using the second neural network;
filter the ranked search results to filter out ranked search results having a match score that fails to satisfy a first threshold;
provide the filtered ranked search results, generated based at least in part on the adjusted match scores, to a first user device, the ranked search results configured to be displayed by the first user device; and
in response to the user selecting a given search results in the filtered ranked search results, cause a download control for a first document to be displayed on the first user device; and
in response to the user activating the download control, download the first document to the first user device.