GRAND universal decoding algorithm is developed to decode any error-correcting code

Sep, 2021 - by CMI


Researchers from Maynooth University in Ireland, Massachusetts Institute of Technology, and Boston University developed a novel silicon device which can crack any form of data transferred over the internet.

Every bit of the data that goes over the internet, from emails to 3D graphics in virtual reality, could be changed by noise, such as electromagnetic intrusion from a Bluetooth device or a microwave. When the data is received at their end point, a decoding algorithm can reverse the adverse effects of the noise and restore the original data.

Researchers from Ireland's Maynooth University, Massachusetts Institute of Technology, and Boston University developed the first-ever silicon chip capable of decoding any code, irrespective of its structure, with high accuracy, by a universal decoding algorithm called as Guessing Random Additive Noise Decoding (GRAND). GRAND enables increased efficiency by removing the requirement for multiple, computationally complex decoders, which could have implementations in virtual and augmented reality, 5G networks, gaming, and network devices that depend on processing a large volume of data with low latency.

GRAND processes by estimating the noise that significantly impacted the message and then determining the original information from the noise pattern. GRAND produces a series of noise patterns in the order in which they are most likely to occur, removes them from the incoming data, and tests whether the generated codeword is in a codebook. Although the noise appears to be random, but has a probabilistic framework that enables the algorithm to predict what it might be.

GRAND chip has three-tiered stages, with the first stage beginning with the simplest potential solutions and progressing to long and more complicated noise sequences in the second and third levels. Each step is self-contained, increasing system throughput while conserving energy. The device can also easily transition between two different codebooks. It has two static random-access memory chips, one of which can decode codewords while the other loads a new codebook and instantly shifts to decoding.

The researcher aims to adjust the silicon chip's structure to improve its efficiency and to test its ability to crack long and complex codes.