Information Technology Reference
In-Depth Information
As an example of the power of receptor degeneracy, Cohen [6] discusses the
example of colour vision in the human eye. The eye possesses millions of colour
receptors called cones of which there are only three types (red, green and blue).
These receptors are degenerate, each responding to broad range of light wave-
lengths that overlap between the different cone types. The human brain, however,
is able to perceive thousands of specific different colours, thus colour specificity is
not encoded by the cones, but achieved via subsequent neuronal firings. Likewise,
Cohen [6] envisages a similar recognition scenario in the immune system.
2.1
Exploiting Degeneracy
The description of degeneracy just presented pitches it as an important, advan-
tageous and powerful property at all levels of biological organisation including
the immune system. Based on this, we have chosen to investigate the property
of degenerate detectors to inspire AIS development. At present there are no in-
stances within the AIS literature where degenerate detectors have been directly
addressed, although degeneracy is an issue that is both being discussed [6,9,11]
and modelled [12] by immunologists. It is clear that incorporating degenerate
detectors into AIS will affect the dynamics of the immune algorithm. Instead of
recognition being the responsibility of a single detector, recognition will emerge
from the collective response of a set of detectors. The assumed benefit of an
AIS with degenerate detectors will be to provide greater scalability and gener-
alisation over existing classifier AIS. Greater scalability can be achieved as the
capacity to discriminate patterns collectively by a set of degenerate detectors
should be greater than by single detectors. Thus, as the number of patterns to
be recognised increases, the number of detectors needed in an AIS with degen-
erate recognition should be less than that of existing AIS. Better generalisation
ability to recognise unseen patterns could be achieved as similar patterns should
produce a similar pattern of response from the set of detectors.
To investigate and exploit degeneracy for the benefit of AIS we follow the
approach previously outlined by us in [1], which advocates the use of the con-
ceptual framework approach [7] to bio-inspired algorithm design. Following this,
and as a first step before building an AIS, we investigate the biology free of any
algorithmic application bias via a process of computational modelling. Based on
the notion that antigen receptors of lymphocytes are degenerate, the aim of this
modelling exercise is to assess the computational impact of lymphocyte antigen
receptor degeneracy on epitope/antigen recognition. This includes investigating
the recognition properties of sets of degenerate receptors when presented with
sets of target antigens. In order to build such a model we first needed to iden-
tify a biological process where recognition by degenerate receptors might take
place. An investigation of suitable immunological literature identified the lymph
nodes as suitable candidate as they are the immune organs where the adaptive
immune response to antigen in the lymph are triggered [13]. Biological details of
the lymph node and T
H
cell activation follow in section 3, which are then used
in the design of an abstract computational model of degeneracy in a lymph node
presented in section 4.
