Robert Boschgmbh

Carnegie Mellon University

Frank R Schmidt

Jeremy Zico Kolter

Eric Wong

Lambros Dalakuras

Jan Hendrik Metzen

Joachim Schenk

Elmar Giehler

Juergen Hirt

Pratyush Maini

Georg Blume

Jan Hendrick Metzen

A method for training an automated learning system includes processing training input with a first neural network and processing the output of the first neural network with a second neural network. The input layer of the second neural network corresponding to the output layer of the first neural network. The output layer of the second neural network corresponding to the input layer of the first neural network. An objective function is determined using the output of the second neural network and a predetermined modification magnitude. The objective function is approximated using random Cauchy projections which are propagated through the second neural network.

Method, apparatus and computer program for generating robust automatic learning systems and testing trained automatic learning systems

A system for training a classification model to be robust against perturbations of multiple perturbation types. A perturbation type defines a set of allowed perturbations. The classification model is trained by, in an outer iteration, selecting a set of training instances of a training dataset; selecting, among perturbations allowed by the multiple perturbation types, one or more perturbations for perturbing the selected training instances to maximize a loss function; and updating the set of parameters of the classification model to decrease the loss for the perturbed instances. A perturbation is determined by, in an inner iteration, determining updated perturbations allowed by respective perturbation types of the multiple perturbation types and selecting an updated perturbation that most increases the loss of the classification model.

Classification robust against multiple perturbation types

In a method for training a first neural network a superposed classification is back-propagated through a second neural network. An output value of the second neural network is utilized to determine whether the input of the first neural network is adversarial.

Method, apparatus and computer program for generating robust automated learning systems and testing trained automated learning systems

A method and device for monitoring a bus system and bus system having at least two users, of which at least one is structured as an authorized user and monitors the data transmission on the bus system, an identifier being transmitted and the identifier being uniquely allocatable to one user for each data transmission on the bus system, the at least one authorized user comparing the identifier with a specified identifier, and if both identifiers are the same taking measures in order to prevent an execution of the data to be transmitted if the data transmission is initiated by a user other than the at least one authorized user.

Method and device for monitoring a bus system and bus system

A system for performing an electronic control in open or closed loop, in particular of functions of a motor vehicle, including at least one electronic control unit and at least one actuator and/or sensor assigned to the control unit, the at least one actuator and/or sensor being controllable via a control unit-actuator/sensor interface, in which variant-determining circuit elements which are arranged outside of the control unit, in particular driver modules are provided between the respective interfaces and the actuator and/or sensor.

System and method for carrying out an electronic control or regulation

For a closed-loop feedback control of a controlled system having a predetermined set course and a control trajectory for use in controlled systems in motor vehicles, an output trajectory of the controlled system is measured for the entire control time in which the control trajectory is applied to the controlled system, an error of the controlled system is detected as a function of predetermined set course and the output trajectory, and a new control trajectory is calculated for each instant of a later control process as a function of the error and the control trajectory.

Growing community of inventors

Stuttgart, Germany

Frank R Schmidt