Towards Unification . . .

This work provides some work towards the unification of the work completed to date as an immunologicalinspired cognitive paradigm. Keywords-Clonal Selection Theory, Acquired Immune System, Artificial Immune System, Cognitive Immune Paradigm I. THE MODELS To date, much effort has been directed towards proposing models and architectures to house the clonal selection theory [1,2,6], and more recently, steps towards the realization of these models [4,5], specifically in the context of a consolidated pathogenic exposure paradigm [3]. The out come of this work may be summarised as follows: Architectures: Three distinct architectures to house clonal selection models, each with their own specific concerns (single repertoire, multiple repertoires, and multiple systems) Components: Many components and component types have been defined. In addition, the tiered abstraction of architectures allows subsumed levels to be used as components at the following level (cell types, tissue types, model types) Processes: Many different processes at each of the three levels of abstraction of the system, each with their own concerns (adaptation, homeostasis, movement, evolution, and immunization sharing) Environments: An environmental paradigm was devised which facilitates stimulus-response adaptation of a situated system (exposure, multiple exposures, and multiple pathogens) Some concerns of the above models that have not yet been reconciled are as follows: (1) a well-defined germinal centre adaptation algorithm, (2) reconciled polyclonal activation and response (an aspect of the cognitive paradigm), (3) a reconciled homeostatic algorithm, (4) a reconciled memory-forming algorithm Some additional concerns which have not yet been integrated into this emerging broader systemenvironment paradigm are as follows: (1) two-phase activation of B and T lymphocytes, (2) reconciled immune network regulation algorithm. II. CLONAL SELECTION IS... A parallel thrust of work has considered clonal selection from a number of different perspectives, including biological, theoretical, and computational learning. Clonal selection is: A. Biological Perspective Immunological Theory: It is a theory that explains the diversity of antibodies via antigenic selection of lymphocyte receptors, which takes into consideration the potential for the production of antibodies that may attack self-tissues. Antiquated Immunological Theory: It is a theory of immunology proposed before the era of molecular science and genetics. It fosters a perspective of discriminated self and nonself, which does not match evidence regarding autoimmunity. It has been augmented with an idiotypic network theory and superseded by danger theory and a cognitive theory of immunology. Clonal Activation Process: It describes a process whereby a relatively high-affinity lymphocyte receptor binds with an antigen, resulting in rounds of cell proliferation, differentiation, mutation, and additional selection. A specialised version of the process applies to both T-lymphocytes and B-lymphocytes. B. Machine Learning Perspective Adaptive Plan: It is a general methodology for a stimulus-response form adaptation from pseudo-random initial conditions, where an environment triggers learning and information is managed through altered receptor (genetic) densities. Learning Algorithm: It is a procedure for populationbased learning through the mechanisms of selection, proliferation, and genetic operators for instantiation and mutation. Cognitive Paradigm: It is an integral component in a broader cognitive paradigm when broader notions such as environment, architecture, and regulatory processes of acquired immunity are taken into consideration. It describes a machine-learning paradigm for investigating cognition and adaptation. C. Others Clonal selection also bares similarity to a number of existent machine learning and artificial intelligence