NON-BIASED PSEUDO RANDOM NUMBER GENERATOR

Patent №

US 5,778,069

Granted

1998-07-07

Filed 1996

Owner

MICROSOFT CORPORATION

AI components

1

hardware

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

08628552

A computer-implemented pseudo random number generator includes an input device to assemble multiple classes of bits from multiple sources into an input bit string. The multiple classes of bits include an internal class of bits from at least one source internal to the random number generator, such as a static bit register which maintains the current state of the generator. The input device also gathers one or more external classes of bits from one or more sources external to the random number generator, such as a machine class of bits which relate to operating parameters of the computer and an application class of bits which relate to execution of an application running on the computer. The input device concatenates the three classes of bits into an arbitrary length input bit string. The random number generator also has a hash computing device which computes an m-bit hash value of the input bit string assembled by the input device. The hash computing device computes the hash value using a hashing function, such as SHA (secure hash algorithm), whereby it is computationally infeasible to derive the concatenated input bit string from the output hash value or intentionally bias the output of the hash function. The SHA is a one-way hash that reduces the 512-bit input bit string to a 160-bit hash value. The hash value becomes the initializing seed for the random number generator. A stream generator (i.e., a stream cipher) is coupled to the hash computing device to receive the hash value. The stream generator uses the hash value as the initializing seed to produce an output bit string of random (or pseudo random) bits.

AI hardwareG06F 7/582G06Q 20/367H04L 9/0662H04L 2209/12H04L 2209/20H04L 2209/56

AI classification

AI hardware0.97
Knowledge representation0.01
Machine learning0.01
Speech0.00
Vision0.00
Natural language0.00
Evolutionary computation0.00
Planning0.00

Ownership

MICROSOFT CORPORATION

assignment · 79460727

Assignors

THOMLINSON, MATTHEW W., SIMON, DANIEL R., YEE, BENNETT

On an employer assignment, the assignors are typically the inventors.

© 2026 NYSGPT2525 LLC