The present invention includes a main device and a backup device. A first backup system element of the main device is used to determine whether a first estimated time is less than a second estimated time. The first estimated time is a time period for directly transmitting pending data. The second estimated time is a time period for transmitting the pending data after the pending data is compressed by a current compression algorithm. If yes, the first backup system element directly transmits the pending data to a second backup system element of the backup device. If not, the first backup system element compresses the pending data by the current compression algorithm, updates the pending data and the current compression algorithm, and re-determines whether the first estimated time is less than the second estimated time. The present invention improves availability of data backup when a network bandwidth is poor.

A method of compressing weights of a neural network includes compressing a weight set including the weights of a the neural network, determining modified weight sets by changing at least one of the weights, calculating compression efficiency values for the determined modified weight sets based on a result of compressing the weight set and results of compressing the determined modified weight sets, determining a target weight of the weights satisfying a compression efficiency condition among the weights based on the calculated compression efficiency values, and determining a final compression result by compressing the weights based on a result of replacing the determined target weight.

A computer-implemented method for storing information into a polynucleotide is provided including using multiple types of nucleotide fragments, wherein each of the nucleotide fragments has an individually different sequence of bases, configuring multiple composite fragments, wherein each of the composite fragments has a set of the nucleotide fragments with different ratios of the nucleotide fragments, and encoding, via an encoder, the information into the composite fragments.

A system and method for low-distortion compaction of floating-point numbers comprising a pre-encoder, a data deconstruction engine, a library manager, a codeword storage, and a data reconstruction engine. A pre-encoder may receive a plurality of data sourcepackets with may contain one or more floating-point numbers and the received data sourcepackets are scanned to identify floating-point numbers and the identified floating-point numbers. Identified floating-point numbers may be pre-encoded into binary string representations which are low-distortion embeddings of real numbers into a Hamming space. The binary string representation may be indexed to indicate it represents a floating-point number before being compacted by a data deconstruction engine and library manager. The pre-encoding of floating-point numbers located within a sourcepacket enables the system to maximize the benefit of the compaction capabilities of the data deconstruction engine.

A processing device, operatively coupled with a memory device, is configured to perform a write operation on a page of a plurality of pages of a data unit of a memory device. The processing device further generates a parity page for data stored in the page of the data unit and associates the parity page with parity data associated with the data unit. Responsive to determining that a first size of the parity data is larger than a first threshold size, the processing device compresses the parity data. Responsive to determining that a second size of the compressed parity data is larger than a second threshold size, the processing device releases at least a subset of the parity data corresponding to a subset of the data that is free from defects.

A method includes defining a lookup table having multiple compressed data values and, for each compressed data value, a range of one or more raw data values associated with the compressed data value. The method also includes compressing raw data using the lookup table. Each raw data value is associated with one of the compressed data values in the lookup table based on a distance ratio involving (i) a distance of the raw data value from a minimum value and (ii) a distance of the raw data value from a maximum value.

The present disclosure pertains to systems and methods to compress an input signal representing a parameter in an electric power system. In one embodiment, a system includes a data acquisition subsystem to receive an input signal comprising a plurality of high-speed representations of electrical conditions. A linear prediction subsystem generates an excitation signal estimate based on the input signal, a plurality of linear prediction coefficients based on the input signal, and an estimated signal based on the excitation signal estimate and the plurality of linear prediction coefficients. An error encoding subsystem may generate an encoding of an error signal based on a difference between the input signal and the estimated signal. A non-transitory computer-readable storage medium may store an encoded and compressed representation of the input signal comprising the excitation signal estimate, the plurality of linear prediction coefficients, and the encoding of the error signal.

The present disclosure relates to a computer implemented method, performed by a wireless device, for reducing data transmission in a communications network by compression of measurements of network characteristics of the communications network, comprising the method steps of: obtaining a trained encoder, of a trained autoencoder, performing a measurements of network characteristics of the communications network, applying the trained encoder to compress the result of the measurements of network characteristics, and transmitting the compressed representation of measurements of network characteristics towards a network entity, whereby the network entity may perform a method comprising the method steps of: obtaining a trained decoder, of the trained autoencoder, receiving a compressed representation of a measurements of network characteristics, and applying the trained decoder to reconstruct the compressed representation of the measurements of network characteristics.

A method of compressing weights of a neural network includes compressing a weight set including the weights of a the neural network, determining modified weight sets by changing at least one of the weights, calculating compression efficiency values for the determined modified weight sets based on a result of compressing the weight set and results of compressing the determined modified weight sets, determining a target weight of the weights satisfying a compression efficiency condition among the weights based on the calculated compression efficiency values, and determining a final compression result by compressing the weights based on a result of replacing the determined target weight.

A method includes defining a lookup table having multiple compressed data values and, for each compressed data value, a range of one or more raw data values associated with the compressed data value. The method also includes compressing raw data using the lookup table. Each raw data value is associated with one of the compressed data values in the lookup table based on a distance ratio involving (i) a distance of the raw data value from a minimum value and (ii) a distance of the raw data value from a maximum value.