Forward and reverse transformations of a model of functional units of a biological system trained by a global model of the systems

A system, comprising:

• a memory that stores computer executable components;
• and
• a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise:
• a parameter identifying component that identifies a plurality of parameters of a multiscale model of a plurality of entities;
• a model generator that generates a low-order model of an entity based on collected data of an operation of the entity, wherein the low-order model of the entity comprises a plurality of transformation modules and a subset of the plurality of parameters;
• and
• a model transformer that trains the low-order model of the entity based on information from the multiscale model of the plurality of entities corresponding to the subset of the plurality of parameters, wherein the transforming the model of the entity results in a first result from the multiscale model of the plurality of entities and a second result from the low-order model of the entity having a relationship that satisfies a defined criterion, given same values used for the subset of the plurality of parameters, wherein the training comprises generating trained parameters for the transformation modules that modulate transformation of outputs of the low-order model to match outputs of the multiscale model according to the defined criterion, wherein the training further comprises an active training simulation to determine active coefficients and a passive training simulation to determine passive coefficients, and wherein the active coefficients and the passive coefficients maximize congruence between the outputs of the low-order model and the outputs of the multiscale model.

A computer program product facilitating training of a low-order model of an entity by a multiscale model of a plurality of entities, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:

• identify, by the processor, a plurality of parameters of the multiscale model of a plurality of entities;
• generate, by the processor, the low-order model of an entity based on collected data of an operation of the entity, wherein the low-order model of the entity comprises a plurality of transformation modules and a subset of the plurality of parameters;
• and
• train, by the processor, the low-order model of the entity based on information from the multiscale model of the plurality of entities corresponding to the subset of the plurality of parameters, wherein the transforming the model of the entity results in a first result from the multiscale model of the plurality of entities and a second result from the low-order model of the entity having a relationship that satisfies a defined criterion, given same values used for the subset of the plurality of parameters, wherein the training comprises generating trained parameters for the transformation modules that modulate transformation of outputs of the low-order model to match outputs of the multiscale model according to the defined criterion, wherein the training further comprises an active training simulation to determine active coefficients and a passive training simulation to determine passive coefficients, and wherein the active coefficients and the passive coefficients maximize congruence between the outputs of the low-order model and the outputs of the multiscale model.