Abstract: A distributed system and method for compressing and restoring data across edge computing devices and cloud infrastructure is disclosed. The system preprocesses raw data at edge computing devices, compresses the data into latent space vectors using distributed encoders within a variational autoencoder spanning edge and cloud components, decompresses the vectors using decoders, and processes them through a resource-aware neural upsampler to generate enhanced reconstructed outputs. The system dynamically adapts compression based on available computing resources and network conditions, while enabling secure distributed processing through homomorphic operations on compressed data. Edge-cloud coordination layers manage data flow, compression parameters, and workload distribution, while maintaining system reliability through intelligent failover handling and resource optimization.

Abstract: A system and method for compressing and restoring data using a hierarchical multi-level autoencoder architecture and correlation network. The system compresses data using a cascade of autoencoders, each focusing on different scales or features, allowing for efficient representation across various resolutions. Data restoration employs a corresponding hierarchical decoder structure and a correlation network, trained on cross-correlated data sets. This approach leverages inter-data relationships at multiple scales, potentially recovering more lost information than traditional single-scale methods. The hierarchical structure adapts to diverse data types, achieving higher compression ratios while maintaining data quality, applicable to fields such as remote sensing, IoT data processing, and multimedia compression.

Abstract: A system and method for enhancing lossy compressed data. The system receives a compressed data stream, decompresses it, and enhances the decompressed data using adaptive neural network models. Key features include data characteristic analysis, dynamic model selection from multiple specialized neural networks, and quality estimation with feedback-driven optimization. The system adapts to various data types and compression levels, recovering lost information without detailed knowledge of the compression process. It implements online learning for continuous improvement and includes security measures to ensure data integrity. The method is applicable to diverse data types, including financial time-series, images, and audio.

Abstract: A system and method for simultaneous compression and encryption of data. The system analyzes input data to determine its properties and creates a transformation matrix based on these properties. Using this matrix, the input data is transformed into a modified distribution, generating a main data stream of transformed data and a secondary stream of transformation information. The main data stream is compressed, and both streams are combined into a single output. The system implements security measures to protect against various attacks, including side-channel vulnerabilities. By using a dyadic distribution algorithm, the system achieves both compression and encryption in a single pass over the data, offering significant efficiency gains. The system can operate in both lossless and lossy modes, providing flexibility for different application requirements. This approach offers a unique solution for data transmission and storage scenarios where both data reduction and security are critical concerns.

Abstract: This invention presents an optimized approach for training and operating Large Language Models (LLMs) using codewords. By converting traditional token-based LLMs to codeword-based systems, the method achieves significant efficiency gains. The process involves tokenizing training data and assigning codewords to tokens. LLMs are then trained and operated using these compact codewords instead of conventional tokens. During operation, prompts are converted to codewords, processed by the LLM, and the outputs are converted back to text. This approach reduces the overall cost of training and operating LLMs by approximately, offering a more efficient solution for large-scale language processing tasks.

Abstract: Codebook data compaction using a universal codebook and mismatch probability estimations to improve entropy encoding methods. Training data sets are analyzed to determine the frequency of occurrence of each sourceblock in the training data sets. A mismatch probability estimate is calculated comprising an estimated frequency at which any given data sourceblock received during encoding will not have a codeword in the codebook. Entropy encoding is used to generate codebooks comprising codewords for data sourceblocks based on the frequency of occurrence of each sourceblock. A “mismatch codeword” is inserted into the codebook based on the mismatch probability estimate to represent those cases when a block of data to be encoded does not have a codeword in the codebook.

Abstract: A system and method for deep learning using a large codeword model with homomorphically compressed and dyadically encrypted data is disclosed. The system preprocesses input data, applies homomorphic-dyadic compression and encryption, tokenizes the compressed data into sourceblocks, and assigns codewords using a codebook. These codewords are processed through a machine learning core, which can be either a conventional transformer-based architecture or a latent transformer core utilizing a variational autoencoder. The system enables secure operations on encrypted data, preserving privacy while allowing complex computations. The processed output is decrypted, decompressed, and translated to match the input modality. A neural upsampler may further enhance the output. The machine learning core is continuously trained using the processed data and additional training data, improving performance over time.

Abstract: A system and method for efficient data storage, transfer, synchronization, and security using automated model monitoring and training. The system analyzes test datasets to detect data drift, retraining encoding and decoding algorithms as needed. New data sourceblocks are created and assigned codewords, compiling an updated codebook for distribution to connected devices. A novel dyadic distribution subsystem simultaneously compresses and encrypts data by transforming input streams into a dyadic distribution. This process generates a compressed main data stream and a secondary stream of transformation information, which are combined into a secure output. The system includes a network device manager for optimizing codebook distribution based on device resource usage. Operating in both lossless and lossy modes, the system offers flexible, efficient, and secure data handling across various network configurations.

Abstract: Image series transformation for optimal compressibility is performed with neural upsampling and error resilience. It incorporates a novel correlation network composed of convolutional layers for feature extraction and a channel-wise transformer with attention to capture complex inter-channel dependencies. The system includes an angle optimizer to further enhance the compressibility of an image and an error resilience subsystem to improve robustness against transmission errors and data loss. The convolutional layers extract multi-dimensional features from the image, while the channel-wise transformer learns global inter-channel relationships. The error resilience subsystem applies forward error correction coding, data partitioning based on importance, and embeds error concealment hints. This hybrid approach addresses both local and global features, mitigates compression artifacts, improves image quality, and enhances data integrity during transmission.

Abstract: A system and method for encrypted data compression, which uses frequency analysis on data blocks within an input data stream to produce a prefix table, representing a first layer of transformation, and which applies a Burrow's-Wheeler transform (BWT) to the data inside the prefix table, representing a second layer of transformation, and which compresses the transformed data. In some implementations, the system and method may further include applying the BWT to a conditioned stream of genomic data, wherein the conditioned stream of data is accompanied by an error stream comprising the differences between the original data and the encrypted data.

Abstract: A system and method for data compaction utilizing distributed codebook encoding to improve entropy encoding methods to account for, and efficiently handle, previously-unseen data in data to be compacted, allow for distributed encoding and decoding capabilities, and allow for parametrized codebook encoding methods. Training data sets are analyzed to determine the frequency of occurrence of each sourceblock in the training data sets. A mismatch probability estimate is calculated comprising an estimated frequency at which any given data sourceblock received during encoding will not have a codeword in the codebook. Further, a codebook and a behavior codebook may both be maintained or altered in a distributed fashion across multiple devices or services, for widespread, or permission-based, or parametrized codebook encoding.

Abstract: A system and methods for upsampling compressed data using a jointly trained Vector Quantized Variational Autoencoder (VQ-VAE) and neural upsampler. The system compresses input data into a discrete latent space using a VQ-VAE encoder, reconstructs the data using a VQ-VAE decoder, and enhances the reconstructed data using a neural upsampler. The VQ-VAE and neural upsampler are jointly trained using a combined loss function, enabling end-to-end optimization. The system allows for efficient compression and high-quality reconstruction of various data types, including financial time-series, images, audio, video, sensor data, and text. The learned discrete latent space can be explored and manipulated using techniques such as interpolation, extrapolation, and vector arithmetic to generate new or modified data samples. The system finds applications in data storage, transmission, analysis, and generation across multiple domains.

Abstract: A system and method for a federated deep learning platform utilizing homomorphically-compressed and encrypted data. The system comprises multiple client devices, each with a local dataset, and a central server hosting a deep learning core. Client devices convert local data into codewords, which are also homomorphically encrypted. The central server processes these encrypted codewords without decryption, preserving data privacy. The platform supports at least two architectural variants: a conventional Transformer trained on codewords, and a Latent Transformer operating on latent space vectors. Both variants eliminate the need for embedding and positional encoding layers. The system aggregates encrypted model updates from clients, enabling collaborative learning while maintaining data confidentiality. Additional features comprise differential privacy implementation and adaptive federated optimization techniques.

Abstract: A system and method for lossy precompression for data compaction using automated model monitoring and training, wherein statistical analyses of test datasets are used to determine if the probability distribution of two datasets are within a pre-determined range, and responsive to that determination new encoding and decoding algorithms may be retrained in order to produce new data sourceblocks, and pre-compression of data prior to processing and statistical analysis allows for the compaction of already compressed data into highly dense formats. The new data sourceblocks may then be processed and assigned new codewords which are compiled into an updated codebook which may be distributed back to encoding and decoding systems and devices.

Abstract: For compressing data, preprocessing operations are performed on raw input data. A discrete cosine transform is performed on the preprocessed data, and multiple subbands are created, where each subband represents a particular range of frequencies. The subbands are organized into multiple groups, where the multiple groups comprise a first low frequency group, a second low frequency group, and a high frequency group. A latent space representation is generated corresponding to each of the multiple groups of subbands. A first bitstream is created based on the latent space representation, and an alternate representation of the latent space is used for creating a second bitstream, enabling multiple-pass techniques for data compression. The multiple bitstreams may be multiplexed to form a combined bitstream for storage and/or transmission purposes.

Abstract: A system and method for highly efficient encoding of data that includes extended functionality for asymmetric encoding/decoding and network policy enforcement. In the case of asymmetric encoding/decoding the original data is encoded by an encoder according to a codebook and sent to a decoder, but the output of the decoder depends on data manipulation rules applied at the decoding stage to transform the decoded data into a different data set from the original data. In the case of network policy enforcement, a behavior appendix into the codebook, such that the encoder and/or decoder at each node of the network comply with network behavioral rules, limits, and policies during encoding and decoding.

Abstract: A system and methods for upsampling of decompressed genomic data after lossy compression using a neural network integrates AI-based techniques to enhance compression quality. It incorporates a novel deep-learning neural network that upsamples decompressed data to restore information lost during lossy compression, taking advantage of cross-correlations between genomic data sets.

Abstract: A system and method for compressing and restoring multi-modal data utilizing a variational autoencoder to enable homomorphic compression techniques is disclosed. Multi-modal input data, comprising at least two different data types, is compressed into a unified latent space using an encoder network of a multi-modal variational autoencoder. Homomorphic operations are performed on compressed data in the latent space. The latent space compressed data is decompressed using a decoder network of the multi-modal variational autoencoder. The system utilizes modality-specific layers and cross-modal attention mechanisms to effectively process diverse data types. The homomorphic operations enable performing computations while the data is in a compressed form, preserving results of those operations in the decompressed output. This approach allows for efficient storage, transmission, and analysis of multi-modal data while maintaining privacy and data integrity across different modalities.

Abstract: A system and method for data storage, transfer, synchronization, and security using automated system efficacy monitoring and model training, wherein statistical analyses of test datasets are used to determine if the probability distribution of two datasets are within a pre-determined range, and responsive to that determination new encoding and decoding algorithms may be retrained in order to produce new data chunklets. The new data chunklets may then be processed and assigned new codewords which are compiled into an updated codebook which may be distributed back to encoding and decoding systems and devices.

Abstract: A system and method for random-access manipulation of compacted data files, utilizing a reference codebook, a random-access engine, a data deconstruction engine, and a data deconstruction engine. The system may receive a data query pertaining to a data read or data write request, wherein the data file to be read from or written to is a compacted data file. A random-access engine may facilitate data manipulation processes by transforming the codebook into a hierarchical representation and then traversing the representation scanning for specific codewords associated with a data query request. In an embodiment, an estimator module is present and configured to utilize cardinality estimation to determine a starting codeword to begin searching the compacted data file for the data associated with the data query. The random-access engine may encode the data to be written, insert the encoded data into a compacted data file, and update the codebook as needed.