US 12,032,688 B2
Method of training a module and method of preventing capture of an AI module
Manojkumar Somabhai Parmar, Ahmedabad (IN); and Adit Jignesh Shah, Ahmedabad (IN)
Assigned to Robert Bosch GmbH, Stuttgart (DE); and Robert Bosch Engineering and Business Solutions Private Limited, Bangalore (IN)
Filed by Robert Bosch GmbH, Stuttgart (DE); and Robert Bosch Engineering and Business Solutions Private Limited, Bangalore (IN)
Filed on Oct. 4, 2021, as Appl. No. 17/493,625.
Claims priority of application No. 2020 4103 3611 (IN), filed on Aug. 6, 2020.
Prior Publication US 2022/0215092 A1, Jul. 7, 2022
Int. Cl. G06F 21/55 (2013.01)
CPC G06F 21/554 (2013.01) [G06F 2221/034 (2013.01)] 8 Claims
OG exemplary drawing
 
1. An AI system comprising (i) an input interface configured to receive input data from at least one user, (ii) an AI module configured to process said received input data, and to generate output data corresponding to said processed input data, and (iii) an output interface configured to send said generated output data from the AI module to the at least one user, the AI system further comprising:
a blocker module configured (i) to detect said received input data is an attack input data, (ii) to manipulate output data of said blocker module, and (iii) to send said manipulated output data from said blocker module to the at least one user; and
a blocker notification module configured to transmit a notification to an owner of said AI system when said attack input data is detected.