One of the major interests in MICREAgents has been the problem of establishing an electronic-chemical code and translation machinery, so that the advantages of sharing electronic and chemical information can be explored. This situation is akin to the major problem of the origin of life with the establishment of the genetic code. Like genetic information, electronic information can be readily manipulated and copied, but its translation into functional chemical information useful for embedded construction processing potentially requires significantly complex machinery and orchestration (cf. the ribosome in cell biology). The object of the work in this area was to apply results from the theory of genetic coding and its evolution to the emergence and control of information transfer between the electronic and chemical domains of lablet behaviour.
The first task was to model the mapping (translation) from digital electronic "memory" states to "operational" chemical reactor states, considering a variety of electronic-chemical information transfer architectures, with a view to identifying those that could enable a hierarchy of phase transitions in the population dynamics. A working model for the dynamics of lablets and their binary associations as gemlabs, including the novel capabilities of interacting autonomous chemistry and electronics was formulated. A detailed description of the elementary states and processes occurring in possible systems of lablets was developed. Essential ingredients of the model are (i) a generic electronic processor based on a simplified microcontroller (ii) a generic chemical processing of chemical coats on electrodes and the local environment (iii) coupling of electronic and chemical signals with docking and undocking of lablets. Docked lablets can communicate with one another both electrically and chemically: they enclose a local chemical environment separate from the global environment and interacting with both lablets involved. The chemical environment can be regarded as the second “information space” into which electronic information is “translated”.
Based on prior joint work on molecular biological systems McCaskill and Wills pursued analysis of the bootstrapping of a translation system in the context of the lablet/gemlab environment; a modeling scheme reflecting the way electronic and chemical information can be encoded and translated in lablet/gemlab systems was developed, considering the overhead in devoting interface resources to support translation, which may be of net value only in certain contexts. A central finding was the observation that the generality of the problems being addressed by the joint information processing system is critical in determining the need for and value of an investment in translation infrastructure. Some of these considerations were documented in a joint journal article. They also connected this work on establishing an electronic-chemical coding system with ideas in the area of reservoir computing. The central idea is that chemical systems based on nonlinear electrochemical and electrokinetic effects can also be sufficiently rich to exhibit the echo state property and thereby serve as a rich dynamical repertoire for learning desired transformations. One special transformation of interest is the communication of information cleanly between lablets: an electronic-chemical-electronic code. Work was initiated to investigate the relationship between the evolution of the genetic code and the evolution of such communication mappings.