Improved methods for coding an ensemble of pulse vectors utilize statistical models (i.e., probability models) for the ensemble of pulse vectors, to more efficiently code each pulse vector of the ensemble. At least one pulse parameter describing the non-zero pulses of a given pulse vector is coded using the statistical models and the number of non-zero pulse positions for the given pulse vector. In some embodiments, the number of non-zero pulse positions are coded using range coding. The total number of unit magnitude pulses may be coded using conditional (state driven) bitwise arithmetic coding. The non-zero pulse position locations may be coded using adaptive arithmetic coding. The non-zero pulse position magnitudes may be coded using probability-based combinatorial coding, and the corresponding sign information may be coded using bitwise arithmetic coding. Such methods are well suited to coding non-independent-identically-distributed signals, such as coding video information.

An arithmetic encoder is provided for converting an event sequence comprised of a plurality of events to an information sequence comprised of at least one information piece, and includes a core engine for receiving an event of the event sequence, and a probability estimate from a probability estimator, and generating zero or more pieces of the information sequence responsive to the received event and the probability estimate by bounding the ratio of events to information pieces. An arithmetic encoder is provided that is capable of constraining a number of events in at least one event sequence as a function of the number of generated information pieces in at least one information sequence. An arithmetic decoder is provided for converting an information sequence comprised of at least one information piece to an event sequence comprised of a plurality of events, and includes a core engine for processing at least one information piece of the information sequence from the sequencer responsive to a probability estimate received from a probability estimator to generate at least one event by accounting for a bounded ratio of events to information pieces in the information sequence.

An arithmetic encoder is provided for converting an event sequence comprised of a plurality of events to an information sequence comprised of at least one information piece, and includes a core engine for receiving an event of the event sequence, and a probability estimate from a probability estimator, and generating zero or more pieces of the information sequence responsive to the received event and the probability estimate by bounding the ratio of events to information pieces. An arithmetic encoder is provided that is capable of constraining a number of events in at least one event sequence as a function of the number of generated information pieces in at least one information sequence. An arithmetic decoder is provided for converting an information sequence comprised of at least one information piece to an event sequence comprised of a plurality of events, and includes a core engine for processing at least one information piece of the information sequence from the sequencer responsive to a probability estimate received from a probability estimator to generate at least one event by accounting for a bounded ratio of events to information pieces in the information sequence.

A system and method for concurrent encryption and lossless compression of data with an algorithm executing on a computer platform. The lossless compression component of the algorithm consists of preprocessing the data with a Burrows-Wheeler transformation followed by an inversion ranking transformation in advance of employing an entropy coder, such as binary arithmetic coder. The frequency vector of the Inversion Ranking transformation is then encrypted and transmitted along with the compressed data with only the frequency vector encrypted. Since the frequency vector is required for decompression, no further encryption of the compressed data is necessary to secure the compressed file.

A network device may receive traffic associated with a network, and may generate a system message based on the traffic. The network device may convert the system message into a binary message, and may compress the binary message to generate a compressed binary message. The network device may provide the compressed binary message to a server device, and the server device may process the compressed binary message, with a decoder, to generate the system message.

One example method includes receiving, by a large language model (LLM), input text to be compressed, defining a size of a rolling window of previous tokens, generated prior to receipt of the input text, that the LLM is permitted to consider in a conditional probability estimate, generating, by the LLM, tokenized text based on the input text, and the tokenized text comprises a sequence of tokens, based on the previous tokens, obtaining a probability mass function of a next token of the sequence, providing the probability mass function as an input to an arithmetic coding (AC) scheme, and assigning, by the AC scheme, a respective binary code to the token with a highest probability as assigned by the LLM.

An system includes a port to receive a number of bits at a first frequency. One or more cells generate a signal for a channel with a channel frequency that is N times greater than the first frequency. The cells transmit at a second frequency that is M times greater than the first frequency but is smaller than the channel frequency. Interface links are coupled between a portion of the input bits of the port and the one or more cells and the portion of the input bits is encoded by thermometer coded T bits such that each one of the T bits is encoded by M repeated parallel bits having a value of a respective T bit. Each interface link includes M interface lines between each T bit and each first cell, and M is smaller than N to reduce the number of interface lines for the T bits.

A method for dealing with respiratory flow data includes steps of: performing integer processing on original flow data to obtain integer flow data; performing difference processing on the integer flow data to obtain difference flow data, the difference flow data having a plurality of difference values in order, the difference values having zero values, positive values, and negative values; when previous one difference value of the zero difference value or one positive difference value is negative, adding a marker value before the difference value that is zero or positive; when previous one difference value of the negative difference value is zero or positive, adding the marker value before the difference value that is negative; performing absolute value processing on the difference flow values to obtain the absolute values; and encoding the absolute values and the marker values to obtain encode flow data. A system using the method is also provided.

Human readable prime number compression (HRPNC) for binary variations, including: generating, in response to a first time threshold being reached during HRPNC of a binary object, one or more variants of the binary object; performing HRPNC on each of the one or more variants of the binary object; and determining, in response to a second time threshold being reached during HRPNC of the one or more variants of the binary object, whether HRPNC of any of the one or more variants has reached a completeness threshold.

Salting binaries for human readable prime number compression (HRPNC), including: detecting, during HRPNC of a binary object, that a time threshold has been reached; generating, in response to the time threshold being reached, one or more salted variants of the binary object by applying, to the binary object, a corresponding salt value; performing HRPNC on each of the one or more salted variants of the binary object; and providing a compressed salted variant.