Real-World Applications - Immunological Computation: Theory and Applications - page 181

Signals
antigen

iDC

Seen Unseen

Signals Signals

smDC

mDC

Match antigen Match antigen

nTC nTC

If Match If Match

aTC nTC

Figure 7.7 Systematic overview of the TLR algorithm.

9. Migrated smDCs and mDCs present their antigen and try and match
nTCs.
10. If an mDC expresses an antigen that matches an nTC receptor, then this
turns the nTC into an activated T cell (aTC) and we have an anomaly.
11. If an smDC expresses an antigen that matches an nTC receptor, then this
kills the nTC to reduce false-positives.
12. Migrated smDCs and mDCs and killed nTCs are replaced with new cells as
per points 4 and 5 (Figure 7.7).

h ere are some similarities between the DCA and TLR algorithms; both perform a
type of temporal correlation between signals and an antigen. However, TLR defi nes
interactions between both T cells and DCs, which is more complex than the single
cell and multiple-signal model employed by the DCA.
Yeom (2007) extended a DT-based approach for network anomaly detection,
where input signals are combined with some information such as data length, name
or identifi cation (ID), or process/program service ID. Here, data is combined with