Systems, Man & Cybernetics - April 2016 - 11

Turf Wars and Buzzwords
The discourse about what CI is (or is not) has not diminished since the arguments about it began in 1992. Chapter
6 of Encyclopedia of Life Support Systems (EOLSS) volume titled Artificial Intelligence [21], authored by Bart
Craenen and A. E. Eiben [22] in 2009, has the title-wait
for it-Computational Intelligence! This chapter has a
pretty accurate and complete recounting of the definitions
of CI offered by Bezdek, Marks, Fogel, and Eberhart that I
reviewed in the previous section. Below are some snippets of their discussion on this topic.
Craenen and Eiben Weigh in on CI
Although used fairly widespread, there is no commonly accepted definition of the term computational
intelligence. Attempts to define, or at least to circumscribe, CI usually fall into one or more of the following categories:
◆ Conceptual treatment of key notions and their roles
in CI
◆ "Relative definition" comparing CI to AI
◆ Listing of the (established) areas that belong to it...The
relationship between computational intelligence and
artificial intelligence has formed a frequently discussed issue during the development of CI. While the
above quote implies they are synonyms, the huge
majority of AI/CI researchers concerned with the subject sees them as different areas, where either
◆ CI forms an alternative to AI
◆ AI subsumes CI
◆ CI subsumes AI. [22]
If you scan the table of contents of the EOLSS volume
Artificial Intelligence [21], you will find among its
chapters these entries: "Ch. 6. Computational Intelligence";
"Ch. 7. Evolutionary Computation"; "Ch. 9. Neural Networks"; "Ch 10. Fuzzy Logic." Evidently the editor of [21]
clearly felt that the presumptive Big 3 that CIS still uses as
the cornerstones of CI, as well as CI itself, are subsets of
AI. This is another data point arguing for the inclusion
structure shown in Figure 1.
Craenen and Eiben suggest the addition of two more
core topics, DNA computing (DNAC) and quantum
computing (QC), to their interpretation of CI, which includes
fuzzy systems computing (FSC), neural networks
computing (NNC), and evolutionary computing (EC). They
compare these five computational styles from three vantage
points: 1) the computational medium, 2) parallelism, and 3)
inspiration from nature. In terms of the computational
medium, they group {FSC, NNC, EC} as silicon-based
computing, while {QC, DNAC} use different environments
for the actual calculations. In a somewhat more
controversial opinion, they call FSC an outlier to the other
four styles in terms of parallelism. In terms of natural
inspiration, they state that FSC and QC do not belong to
natural computation-that is, having been inspired by
natural processes. I find this a bit puzzling, since the entire

field of fuzzy systems is based on the idea of representing
natural language computationally. Puzzling, but so what?
This is a really good chapter to read, and it provokes a lot of
thought about what CI is and is not. I recommend it.
On the other hand, have a look at the table of contents
for the EOLSS volume Computational Intelligence,
edited by Hisao Ichibuchi [23], in Table 2. To be sure, this
volume is largely composed of topics related to the Big 3,
but it would be pretty unfair to hold the CIS's hand to the
fire for departing from its technical roots of 30 years ago.
In the main the topics shown still represent the Big 3. For
example, I regard support vector machines as a subfield of
CNNs, but you may wish to call it something else. Swarm
intelligence? It's a low-level optimization technique that
seems to fit in the EC domain.
There are also some fields represented by publications
under the IEEE CIS banner that don't have a sufficiently large
current support set to warrant inclusion in [23]: virtual reality,
financial engineering, autonomous mental development,
game theory, bioinformatics, information forensics, and so
forth. That's fine-topics come and go just like buzzwords, or
cuisines. We see additional topics that reflect technological
evolution in both AI and CI. For example, Chapter 8 in [21]

Table 2. The table of contents for the
book Computational Intelligence [23].
1. History of Computational Intelligence (this article, long
form, somewhat dated now)
2. History and Philosophy of Neural Networks
3. Recurrent Neural Networks
4. Adaptive Dynamic Programming and Reinforcement
Learning
5. Associative Learning
6. Kernel Models and Support Vector Machines
7. The Genesis of Fuzzy Sets and Systems-Aspects in Science
and Philosophy
8. Design and Tuning of Fuzzy Systems
9. Fuzzy Data Analysis
10. Introduction to Interval Type-2 Fuzzy Logic Systems
11. Rough Set Approximations-A Concept Analysis Point of View
12. Evaluating the Evolutionary Algorithms-Classical
Perspectives and Recent Trends
13. General Framework of Evolutionary Computation
14. Evolutionary Multiobjective Optimization
15. Memetic Algorithms
16. Swarm Intelligence
17. Artificial Immune Algorithms in Learning and Optimisation
18. Hybrid Computational Intelligence
19. Computational Intelligence and Medical Applications
20. Computational Intelligence and Smart Grids
21. Computational Intelligence and Computational Systems
Biology
22. Computational Neuroscience
23. Neuromorphic Engineering
24. Brain-Machine Interface

Ap ri l 2016

IEEE SyStEmS, man, & CybErnEtICS magazInE

Table of Contents for the Digital Edition of Systems, Man & Cybernetics - April 2016