# US 2021/0241074 A1 · System for generating synthetic digital data for data multiplication > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2021/0241074 A1 - Document ID: US20210241074A1 - Status: Published application - Kind: A1 - Application number: 16/778,813 - Filed: 2020-01-31 - Published: 2021-08-05 - Earliest claimed: 2020-01-31 - Assignees: Rohde & Schwarz GmbH & Co. KG - Applicants: Rohde & Schwarz GmbH & Co. KG - Inventors: Timo MAYER; Mikhail VOLIANSKII - Web page: https://patentgrep.com/patent/US20210241074A1 ## CPC - G06N 3/045: Combinations of networks - G06N 3/0455: Auto-encoder networks; Encoder-decoder networks - G06N 3/0475: Generative networks - G06N 3/0495: Quantised networks; Sparse networks; Compressed networks - G06N 3/08: Learning methods - G06N 3/09: Supervised learning - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract The invention relates to a system for generating synthetic digital data, comprising a receiver configured to receive a measured signal, in particular an RF signal, a converter configured to convert the measured signal to digital data representing signal characteristics of the measured signal, a trainable neural network encoder and a trainable neural network decoder, wherein, during a training routine, the neural network encoder is configured to receive the digital data and to generate a compressed representation of the digital data, and the neural network decoder is configured to generate a reconstruction of the digital data based on the compressed representation, and wherein the trained neural network decoder is configured to receive random or pseudorandom data and to generate synthetic digital data representing measured signal characteristics based on the random or pseudorandom data. ## Claims ### Claim 1 A system for generating synthetic digital data, comprising: a receiver configured to receive a measured signal, in particular an RF signal, a converter configured to convert the measured signal to digital data representing signal characteristics of the measured signal, a trainable neural network encoder and a trainable neural network decoder, wherein, during a training routine, the neural network encoder is configured to receive the digital data and to generate a compressed representation of the digital data, and the neural network decoder is configured to generate a reconstruction of the digital data based on the compressed representation, and wherein the trained neural network decoder is configured to receive random or pseudorandom data and to generate synthetic digital data representing measured signal characteristics based on the random or pseudorandom data. ### Claim 2 (depends on 1) The system of claim 1, comprising a data storage, wherein the data storage is configured to store the digital data and the generated synthetic digital data, or wherein the data storage is configured to only store the generated synthetic digital data. ### Claim 3 (depends on 2) The system of claim 2, wherein the data storage comprises a physical memory and/or a cloud storage. ### Claim 4 (depends on 1) The system of claim 1, comprising a transmitter, wherein the transmitter is configured to transmit the digital data and the generated synthetic digital data to a receiver device, or wherein the transmitter is configured to only transmit the generated synthetic digital data to the receiver device. ### Claim 5 A method for generating synthetic digital data, comprising the step of: feeding random or pseudorandom data to a trained neural network decoder, wherein the trained neural network decoder is configured to generate synthetic digital data representing measured signal characteristics based on the random or pseudorandom data, wherein the trained neural network decoder is obtainable by: a) receiving a measured signal, in particular an RF signal, at a receiver, b) converting the measured signal to digital data representing signal characteristics of the measured signal, c) feeding said digital data to a neural network encoder, wherein the neural network encoder is configured to generate a compressed representation of the digital data, and d) feeding the compressed representation of the digital data to the neural network decoder, wherein the neural network decoder is configured to generate a reconstruction of the digital data based on the compressed representation. ### Claim 6 (depends on 5) The method of claim 5, wherein the digital data and the generated synthetic digital data are stored in a data storage, or wherein only the generated synthetic digital data is stored in the data storage. ### Claim 7 (depends on 5) The method of claim 5, wherein the digital data and the generated synthetic digital data are transmitted to a receiver device, or wherein only the generated synthetic digital data is transmitted to the receiver device. ### Claim 8 (depends on 5) The method of claim 5, wherein the measured signal and/or the digital data comprises RF channel measurement data, RF test drive data, RF field test data, time data and/or position data. ### Claim 9 (depends on 5) The method of claim 5, wherein the trained neural network decoder is further obtainable by: e) comparing the reconstructed digital data generated by the neural network decoder to the corresponding digital data fed to the neural network encoder, f) determining a loss function, representing a loss between the digital data and the corresponding reconstructed digital data, and g) adapting the neural network decoder to based on the loss function. ### Claim 10 (depends on 5) A test method for a physical entity, in particular a communication channel or a measurement device, comprising feeding the entity with the synthetic data generated according to the method of claim 5, and performing a real world measurement as to characteristics of the entity. ### Claim 11 (depends on 5) Use of synthetic digital data generated according to the method of claim 5 in a test method for a physical entity. ### Claim 12 (depends on 5) A computer program comprising a program code for performing the method of claim 5 when executed on a computer.