Model calculating unit and control unit for calculating a neural layer of a multilayer perceptron model

Patent metadata

Field	Value
Publication number	US-11645499-B2
Application number	US-201716466814-A
Country	US
Kind code	B2
Filing date	Sep 4, 2017
Priority date	Sep 7, 2016
Publication date	May 9, 2023
Grant date	May 9, 2023

Abstract

A model calculating unit for calculating a neural layer of a multilayer perceptron model having a hardwired processor core developed in hardware for calculating a definitely specified computing algorithm in coupled functional blocks. The processor core is designed to calculate, as a function of one or multiple input variables of an input variable vector, of a weighting matrix having weighting factors and an offset value specified for each neuron, an output variable for each neuron for a neural layer of a multilayer perceptron model having a number of neurons, a sum of the values of the input variables weighted by the weighting factor, determined by the neuron and the input variable, and the offset value specified for the neuron being calculated for each neuron and the result being transformed using an activation function in order to obtain the output variable for the neuron.

First claim

What is claimed is:

1. A model calculating unit for calculating a layer of a multilayer perceptron model, comprising:
a hardwired processor core configured in hardware to calculate a definitely specified computing algorithm in coupled functional blocks, the processor core being configured to calculate an output variable for each neuron for a neural layer of the multilayer perceptron model having a number of neurons as a function of: (i) one or multiple input variables of an input variable vector, (ii) a weighting matrix having weighting factors, and (iii) an offset value specified for each of the neurons, a sum of the values of the input variables, weighted by a weighting factor determined by the neuron and the input variable, and the offset value specified for the neuron being calculated for each neuron, and a result being transformed using an activation function to obtain the output variable for the neuron;
wherein the model calculating unit includes at least one of the following features (a)-(b):
(a) the functional blocks include a state machine hardwired such that the state machine is configured to at least one of: (1) select between implementing the multilayer perceptron model and at least one alternative model based on a first received input variable, and (2) select between implementing different activation functions with the multilayer perceptron model based on a second received input variable; and
(b) the processor core is configured to base the calculation of the output variable on distance values that are a function of respective distances of respective neuron values of the neurons from the values of the input variable vector, with a weighting of the distance values by respective length scales assigned to respective ones of the neurons.

2. The model calculating unit as recited in claim 1 , wherein the processor core includes the state machine, a memory to store the one or multiple input variables of the input variable vector, the weighting matrix, the offset values specified for each neuron, and the output variables for each neuron, and one or multiple computing operation blocks, the computing operating blocks including a MAC block, and an activation function calculation block.

3. The model calculating unit as recited in claim 1 , wherein the processor core is developed in a surface area of an integrated chip.

4. The model calculating unit as recited in claim 1 , wherein the processor core is configured to base the calculation of the output variable on the distance values that are a function of the respective distances of the respective neuron values of the neurons from the values of the input variable vector, with the weighting of the distance values by the respective length scales assigned to the respective ones of the neurons.

5. The model calculating unit as recited in claim 4 , wherein the distance values are quadratic distance values.

6. The model calculating unit as recited in claim 4 , wherein the distance values are squares of the respective distances.

7. The model calculating unit as recited in claim 1 , wherein the functional blocks include the state machine, the state machine being hardwired such that the state machine is configured to select between the implementing of the multilayer perceptron model and the at least one alternative model based on the first received input variable.

8. The model calculating unit as recited in claim 7 , wherein the state machine is hardwired such that the state machine is configured to, when the multilayer perceptron model is selected instead of the at least one alternative model, additionally select between the implementing of the different activation functions with the multilayer perceptron model based on the second received input variable.

9. The model calculating unit as recited in claim 8 , wherein the at least one alternative model includes at least one of a Gaussian process model and a Radial Basis Function (RBS) model.

10. The model calculating unit as recited in claim 8 , wherein the processor core is configured to base the calculation of the output variable on the distance values that are a function of the respective distances of the respective neuron values of the neurons from the values of the input variable vector, with the weighting of the distance values by the respective length scales assigned to the respective ones of the neurons.

11. The model calculating unit as recited in claim 7 , wherein the at least one alternative model includes at least one of a Gaussian process model and a Radial Basis Function (RBS) model.

12. The model calculating unit as recited in claim 11 , wherein the different activation functions, between which the state machine is configured to select, include a kink function, a sigmoid function, a hyperbolic tangent function, and a linear function.

13. The model calculating unit as recited in claim 7 , wherein the processor core is configured to base the calculation of the output variable on the distance values that are a function of the respective distances of the respective neuron values of the neurons from the values of the input variable vector, with the weighting of the distance values by the respective length scales assigned to the respective ones of the neurons.

14. The model calculating unit as recited in claim 1 , wherein the functional blocks include the state machine, the state machine being hardwired such that the state machine is configured to select between the implementing of the different activation functions with the multilayer perceptron model based on the second received input variable.

15. The model calculating unit as recited in claim 14 , wherein the processor core is configured to base the calculation of the output variable on the distance values that are a function of the respective distances of the respective neuron values of the neurons from the values of the input variable vector, with the weighting of the distance values by the respective length scales assigned to the respective ones of the neurons.

16. A control unit, comprising:
a microprocessor; and
one or multiple model calculating units for calculating a layer of a multilayer perceptron model, each of the model calculating units including a hardwired processor core configured in hardware to calculate a definitely specified computing algorithm in coupled functional blocks, the processor core being configured to calculate an output variable for each neuron for a neural layer of the multilayer perceptron model having a number of neurons as a function of: (i) one or multiple input variables of an input variable vector, (ii) a weighting matrix having weighting factors, and (iii) an offset value specified for each of the neurons, a sum of the values of the input variables, weighted by a weighting factor determined by the neuron and the input variable, and the offset value specified for the neuron being calculated for each neuron, and a result being transformed using an activation function to obtain the output variable for the neuron;
wherein the model calculating unit includes at least one of the following features (a)-(b):
(a) the functional blocks include a state machine hardwired such that the state machine is configured to at least one of: (1) select between implementing the multilayer perceptron model and at least one alternative model based on a first received input variable, and (2) select between implementing different activation functions with the multilayer perceptron model based on a second received input variable; and
(b) the processor core is configured to base the calculation of the output variable on distance values that are a f

Assignees

• Bosch Gmbh Robert

Inventors

• Guntoro Andre
• Kloppenburg Ernst
• Markert Heiner
• Schiegg Martin

Classifications

• G06N3/0499

  Feedforward networks · CPC title

• G06N3/063 Primary

  using electronic means · CPC title

• F02D41/266

  the computer being backed-up or assisted by another circuit, e.g. analogue · CPC title

• G06N3/048 Primary

  Activation functions · CPC title

• G05B13/027

  using neural networks only · CPC title