Robert bosch gmbh (20240211735). ERROR-PROOF INFERENCE CALCULATION FOR NEURAL NETWORKS simplified abstract
Contents
ERROR-PROOF INFERENCE CALCULATION FOR NEURAL NETWORKS
Organization Name
Inventor(s)
Christoph Schorn of Leonberg (DE)
Leonardo Luiz Ecco of Stuttgart (DE)
Andre Guntoro of Weil Der Stadt (DE)
Sebastian Vogel of Schaidt (DE)
ERROR-PROOF INFERENCE CALCULATION FOR NEURAL NETWORKS - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240211735 titled 'ERROR-PROOF INFERENCE CALCULATION FOR NEURAL NETWORKS
Simplified Explanation:
The patent application describes a method for operating a hardware platform to calculate inferences in a convolutional neural network. The method involves convolving an input matrix with multiple convolution kernels to generate two-dimensional output matrices, summing these kernels to create a control kernel, convolving the input matrix with the control kernel to produce a control matrix, and comparing elements of the control matrix with corresponding elements in the output matrices to ensure accuracy.
- The method involves convolving input data with convolution kernels to generate output matrices.
- Summing the convolution kernels to create a control kernel for further convolution.
- Comparing elements of the control matrix with elements in the output matrices to verify accuracy.
Potential Applications:
This technology can be applied in various fields such as image recognition, natural language processing, and autonomous driving systems where convolutional neural networks are used for inference calculations.
Problems Solved:
This technology addresses the need for accurate inference calculations in convolutional neural networks by verifying the correctness of output matrices through control calculations.
Benefits:
The benefits of this technology include improved accuracy in inference calculations, enhanced reliability of convolutional neural networks, and potential for real-time error detection and correction.
Commercial Applications:
Title: Enhanced Inference Calculation Method for Convolutional Neural Networks
This technology can be utilized in industries such as healthcare for medical image analysis, retail for customer behavior analysis, and security for facial recognition systems. The market implications include improved efficiency, accuracy, and reliability in neural network applications.
Prior Art:
Prior research in the field of convolutional neural networks and hardware acceleration methods for inference calculations can provide valuable insights into the development and implementation of this technology.
Frequently Updated Research:
Researchers are constantly exploring new techniques and algorithms to optimize inference calculations in convolutional neural networks, which may impact the future advancements of this technology.
Questions about Convolutional Neural Network Inference Calculation:
1. How does this method improve the accuracy of inference calculations in convolutional neural networks? 2. What are the potential real-time applications of error detection and correction in neural network calculations?
Original Abstract Submitted
a method for operating a hardware platform for the inference calculation of a convolutional neural network. in the method: an input matrix having input data of the neural network is convolved by the acceleration module with a plurality of convolution kernels, so that a multiplicity of two-dimensional output matrices results; the convolution kernels are summed elementwise to form a control kernel; the input matrix is convolved by the acceleration module with the control kernel, so that a two-dimensional control matrix results; each element of the control matrix is compared with the sum of the elements corresponding thereto in the output matrices; if this comparison yields a deviation for an element of the control matrix, then in response it is checked, with at least one additional control calculation, whether an element of at least one output matrix corresponding to this element of the control matrix was correctly calculated.
