Deep neural network implementation for concatenated codes

A computer-implemented method to perform operations associated with a storage medium, the computer-implemented method comprising:

• receiving, for each component of a plurality of components of product code, a set of input log-likelihood ratios (LLRs) of a component of the plurality of components;
• determining, for each component, an extrinsic estimation function based on a set of features of the set of input LLRs;
• analyzing, for each component, the extrinsic estimation function to obtain a plurality of extrinsic LLR values;
• mapping the plurality of extrinsic LLR values to an input LLR of the set of input LLRs;
• and
• outputting, for each component, a plurality of updated LLR values based on the mapping.

A non-transitory processor-readable medium containing processor-readable instructions, such that, when executed by one or more processors, performs operations associated with a storage medium by:

• receiving, for each component of a plurality of components of product code, a set of input log-likelihood ratios (LLRs) of a component of the plurality of components;
• determining, for each component, an extrinsic estimation function based on a set of features of the set of input LLRs;
• analyzing, for each component, the extrinsic estimation function to obtain a plurality of extrinsic LLR values;
• mapping the plurality of extrinsic LLR values to an input LLR of the set of input LLRs;
• and
• outputting, for each component, a plurality of updated LLR values based on the mapping.