Information Technology Reference
In-Depth Information
The quality of work can be unclear if the terminology used to describe it is
over-inflated. Sometimes such terminology is used to avoid having to define terms
properly. A
hypernet
, for example, sounds much more powerful than a network;
but who knows if there is really a difference. Researchers use such terminology to
make cloudy, big-picture claims that are rarely justified by their actual outcomes.
Terms that in common usage describe aspects of cognition or consciousness, such
as “intelligent” or “belief”, or even “aware”, are particularly slippery. They sound
like ordinary concepts we are all familiar with. But in their common usage they are
not well defined; and when terms are borrowed from common usage their meaning
changes. These terms anthropomorphize the computational behaviour to create a
sense of specialness or drama, when in fact what is being described may well be
highly mechanical and deterministic, and possibly isn't very interesting. This is a
form of the renaming fallacy noted earlier. Thus, while we might have an impression
of what the author means when they claim that a system is intelligent, that impression
is vague. Successful science is not built on vagueness.
Aparticular example is thewidelymisused term“semantic”, which, strictly speak-
ing, concerns the meaning of a concept, as distinct from its syntax or representation.
But computers are machines for processing representations: enriching the represen-
tation by, say, addition of further descriptors does not “bridge the semantic gap”.
At best, it shifts the problem, from one of computing in the absence of descriptors
to one of creating and then making use of descriptors. For example, a text indexing
technique that “maps terms to concepts, allowing semantic retrieval” might be no
more than a trivial function in which an ontology is used to map words, correctly
or otherwise, to labels; retrieval then proceeds as usual, but with labels instead of
terms as queries. An additional resource (a dictionary) has been introduced into the
process, but the method isn't semantic, and certainly isn't particularly intelligent.
Some science is not simply weak, but can be described as pseudoscience.
Inevitably, some claimed achievements are delusional or bogus. Pseudoscience is
a broad label covering a range of scientific sins, from self-deception and confusion
to outright fraud. A definition is that pseudoscience is work that uses the language
and respectability of science to gain credibility for statements that are not based
on evidence that meets scientific standards. Much pseudoscience shares a range of
characteristics: the results and ideas don't seem to develop over time, systems are
never quite ready for demonstration, the work proceeds in a vacuum and is unaffected
by other advances, protagonists argue rather than seek evidence, and the results are
inconsistent with accepted facts. Often such work is strenuously promoted by one
individual or a small number of devotees while the rest of the scientific community
ignores it.
2
2
An example of pseudoscience in computing are schemes for high-performance video compression
that promised delivery of TV-quality data over low-bandwidth modems. In the 1990s, the commer-
cial implications of such systems were enormous, and this incentive created ample opportunities
for fraud. In one case, for example, millions of dollars were scammed from investors with tricks
such as hiding a video player inside a PC tower and hiding a network cable inside a power cable.
Yet, skeptically considered, such schemes are implausible. For example, with current technol-
ogy, even a corner of a single TV-resolution image—let alone 25 frames—cannot be compressed into
