Back to results
US 2023/0068246 A1Published application
Database utilizing spatial probability models for data compression
OwnerAdam Jeffriesinventor
Life of this patent
Estimating the term…
Abstract
A method, article comprising machine-readable instructions and apparatus that processes data systems for encoding, decoding, pattern recognition/matching and data generation is disclosed. State subsets of a data system are identified for the efficient processing of data based, at least in part, on the data system's systemic characteristics.
Claims17
A system comprising:
- a processing device;
- computerized system a non-transitory, tangible medium that stores instructions that when executed by the processing device cause the system to perform operations comprising at least:
- matching systemic patterns of a data system to a spatial probability model, and/or
- forecasting a probability value corresponding to a future interval and/or a previous interval comprised by a time dimension comprised by a space of the spatial probability model, wherein the spatial probability model is configured to reduce memory utilization by incorporating a sequence of probability distributions that vary along intervals in the time dimension comprised by the space of the spatial probability model, the time dimension corresponding to at least a length of a monotonic output sequence;
- utilizing the spatial probability model to determine a probability of a state comprised by a state superset of at least one of:
- a data unit, a first monotonic output sequence, a data system, or a systemic characteristic of the data system,
- by at least utilizing the spatial probability model to perform:
- utilizing a referencing system or a referencing function that determines a correspondence between a current second data unit state of a second data unit comprised by a second monotonic output sequence and a probability distribution of possible states of the data unit comprised by the second monotonic output sequence, the probability distribution of possible states of the data unit comprised by the sequence of probability distributions that vary along intervals in the time dimension comprised by the space of the spatial probability model, the probability distribution of the possible states of the data unit corresponding to an interval in the time dimension, the second monotonic output sequence based, at least in part, on receiving an ordered sequence of data units comprised by a second data system and utilizing a sequence function to generate the second monotonic output sequence based on the ordered sequence of data units,
- determining a spatial probability model output based, at least in part, on a spatial probability model value, enabling a determination of a probability of the current second data unit state of the second data unit comprised by the second monotonic output sequence based, at least in part, on a number of correspondences determined by the referencing system or the referencing function between the probability distribution of the possible states of the data unit and the current data unit state of the data unit comprised by the monotonic output sequence;
- using the spatial probability model output based, at least in part, on the spatial probability model value to perform at least one of:
- relatively reduce memory utilization and utilization of computational resources when utilizing the spatial probability model to match systemic characteristic patterns of an incoming data system with systemic characteristics described by the spatial probability model and/or, to forecast by a forecaster a forecasted data system, the incoming data system and the forecasted data system corresponding to at least one of:
- an encoded message comprising symbols, a document comprising graphemes or alphanumeric symbols, a digital picture comprising pixels, a waveform file comprising samples, a video file comprising picture frames, a digital scene comprising points, polygons or voxels, a pseudorandom file comprising data with pseudorandom characteristics, or a forecasted ordered sequence of data units corresponding to predicted data associated with a likelihood of occurring at the future interval or the previous interval comprised by the time dimension.
9 more elements
A method comprising:
- utilizing, by a computerized system comprising a data processor comprising hardware, a spatial probability model, that relatively reduces memory utilization, utilization of computational resources and reduces information distortion by comprising a sequence of probability distributions that vary along intervals in a time dimension corresponding to at least a length of a monotonic output sequence to determine a probability of a state comprised by a state superset of at least one of:
- a data unit, the monotonic output sequence, a data system or a systemic characteristic of the data system by at least utilizing the spatial probability model to perform:
- determining a point comprised by a space of the spatial probability model by utilizing a referencing system or a referencing function that determines a correspondence between the point and a current second data unit state of a second data unit comprised by a second monotonic output sequence, the second monotonic output sequence based, at least in part, on receiving an ordered sequence of data units comprised by a second data system and utilizing a sequence function to generate the second monotonic output sequence based on the ordered sequence of data units, the second data system comprising mutually dependent units of data, the point corresponding to at least:
- a data unit state in a state dimension, the state dimension comprising possible data unit states of the data unit comprised by the monotonic output sequence and, a probability distribution of the possible states of the data unit comprised by the monotonic output sequence, the probability distribution of the possible states of the data unit comprised by the sequence of probability distributions that vary along intervals in the time dimension and, an interval in the time dimension, corresponding to at least the length of the monotonic output sequence;
- and
- determining a spatial probability value linked to the point comprised by the space of the spatial probability model, the spatial probability value enabling a determination of a probability of a correspondence between the point and a current second data unit state of the second data unit comprised by the second monotonic output sequence, the spatial probability value associated with probability distributions that vary along intervals in the time dimension, resulting in a relative reduction in memory utilization, utilization of processing resources, and/or information distortion in determining the probability of the state comprised by the state superset of at least one of:
- the data unit, the monotonic output sequence, the data system or the systemic characteristic of the data system;
- determining, by the data processor, a spatial probability model output based, at least in part, on the spatial probability model value;
- using the spatial probability model output based, at least in part, on the spatial probability model value to perform at least one of:
- relatively reducing memory utilization and utilization of computational resources when utilizing the spatial probability model to match systemic characteristic patterns of an incoming data system with systemic characteristics described by the spatial probability model and/or, to forecast by a forecaster a forecasted data system, the incoming data system and the forecasted data system corresponding to at least one of:
- an encoded message comprising symbols, a document comprising graphemes or alphanumeric symbols, a digital picture comprising pixels, a waveform file comprising samples, a video file comprising picture frames, a digital scene comprising points, polygons or voxels, a pseudorandom file comprising data with pseudorandom characteristics, or a forecasted ordered sequence of data units corresponding to predicted data associated with a likelihood of occurring at a future interval or a previous interval comprised by the time dimension.
8 more elements
Dependencies are inferred from the claim text.