What is an Artificial Immune System (AIS)?
An artificial immune system uses biological immune systems to implement algorithms or technologies that systemic targets. This may involve the mathematical and computer modelling of immune systems or the abstraction of certain principles related to immunology in algorithms.
History of Artificial Immune System
The artificial immune systems emerged in the mid-1980s with articles written by Farmer, Packard, and Perelson (1986) and Bersini and Varela (1990) on immune networks. However, it wasn’t until the mid-1990s that AIS became a field in its own right.
Forrest et al. (on negative selection) and Kephart et al. published their first articles on AIS in 1994, and Dasgupta conducted extensive studies on negative selection algorithms. Hunt and Cooke began working on immune network models in 1995; Timmis and Neal continued this work and improved.
New ideas are currently being explored in artificial immune systems, such as the theory of danger and algorithms inspired by the innate immune system. However, some believe that these new ideas do not yet offer a true “new” summary, in addition to the existing AIS algorithms.
However, this is hotly debated, and debate is currently one of the main driving forces behind the development of AIS. Other recent developments involve exploring degeneracy in AIS models, motivated by its hypothetical role in open learning and evolution.
Originally, the artificial immune systems sought to find efficient abstractions of the processes found in the immune system. Still, more recently, it has been interested in the modelling of biological processes and the application of immune algorithms to bioinformatics problems.
In 2008, Dasgupta and Nino published a manual on immunological calculus that presents a collection of up-to-date work related to immunity-based techniques and describes a wide variety of applications.
What are the Uses of Artificial Immune System?
Like artificial neural networks, artificial immune systems can learn new information, remember previously learned information, and perform pattern recognition in a highly decentralized way. This volume provides an overview of the immune system from a computational perspective.
1. Technical
Current techniques are inspired by specific immunological theories that explain the function. And also, the behaviour of the adaptive immune system in mammals.
2. Clonal Selection Algorithm
A class of algorithms enthused by the clonal selection philosophy of acquired immunity explains how B. And also, T cells improve their response to antigens over time, called affinity maturation. These algorithms focus on Darwinian attributes of the theory.
The affinity of antigen-antibody interactions inspires selection, reproduction inspires by cell division, and variation inspires by somatic hypermutation. And also, Clonal selection algorithms most often apply to optimization and pattern recognition domains, resembling parallel escalation. And also, the genetic algorithm without the recombination operator.
3. Negative Selection Algorithm
Inspired by the positive and negative selection processes that occur during maturation of T cells in the thymus, called T cell tolerance. Negative selection refers to the identification and suppression (apoptosis) of autoreactive cells. And also, T cells can select and attack the tissues of the self.
This class of algorithms generally uses pattern recognition. And also, classification problem areas where the problem space model complements the available knowledge. For example, in an anomaly detection domain, the algorithm prepares a set of exemplary pattern detectors formed on normal (non-abnormal) patterns that model. And also, detect invisible or abnormal patterns.
4. Immune Network Algorithms
Algorithms inspired by Niels Kaj Jerne’s idiotypic network theory describe the immune system’s regulation by anti-idiotypic antibodies (antibodies that select other antibodies). This class of algorithms focuses on the network graph structures involved where antibodies (or antibody-producing cells) represent nodes.
The learning algorithm involves the growth or pruning of edges between nodes as a function of affinity (similarity in space representation of the problems). Immune network algorithms have used in the grouping, data visualization, control and optimization. And also, share properties with artificial neural networks.
5. Dendritic Cell Algorithms
The Dendritic Cell Algorithm (DCA) an example of an immune system-inspired algorithm developed using a multiscale approach. This algorithm base on an abstract dendritic cell (DC) model. And also, DCA summarize and implemented through examining and modelling various aspects of the DC function.
DCA from the molecular networks presents within the cell to the behaviour exhibited by a population of cells as a whole. Within the DCA, information granulate at different layers, achieved through multi-scale processing.
Conclusion
Artificial Immune Systems (AIS) concern with the abstraction of the structure and function of the immune system from computer systems. And studying the application of these systems to solve computer problems in mathematics, engineering and information technology.
AIS a subfield of biologically inspired computing and natural computing, with interests in machine learning. And also, belonging to the broader field of artificial intelligence. Artificial immune systems (AIS) adaptive systems inspired by theoretical immunology and observed immune functions, principles, and models applied to problem-solving.