Biomedical Engineering Reference
In-Depth Information
Although it is no longer necessary to visit a library in person to read a
scholarly work, so much material now exists that making sense of
what is relevant is becoming increasingly diffi cult. In the life sciences
alone, a new article is published every two minutes [1], and the number
of articles has been growing exponentially at around 4% per year for
the past 20 years [2]. In 2011, for example, over 14 000 papers were
published relating in some way to Human Immunodefi ciency Virus
(admittedly, this is a much broader topic than any one researcher would
care to explore in detail; nevertheless, it gives some indication of the scale
of the problem) [3]. If we add to the peer-reviewed literature all the
information that is deposited in databases, or written informally on
personal blogs, matters get more complex still. The problem has been
characterized, at a broad level, as being one of 'fi lter failure' [4], that is
it's not that there is too much information per se - as, surely, lots of
information has to be a good thing, and there has always been more
available than we are able to consume - but, rather, that we are lacking
mechanisms to convert information into meaningful knowledge, to fi nd
what is relevant to our particular needs, and to make connections between
the disjoint data.
The problem is particularly acute in the life sciences. As pharmaceutical
companies race to fi nd new drugs, it has become clear that mastery of the
literature is at least as important as mastery of the test tube or the
sequencing machine. It has been documented that some failures of
pharmaceutical company projects - which are costly on a mind-boggling
scale, and typically result in failure - could have been avoided if scientists
within the company had simply been able to fi nd the relevant scientifi c
article beforehand [5]. This inability to fi nd the 'right article' is not for
the want of trying: pharmaceutical companies employ large teams
of scientists to scour the literature for just such information; it is
simply that, even with the best tools and the most profi cient scientists,
there is more information 'out there' than can currently be processed in a
timely way.
The danger of characterizing this as simple 'fi lter failure' is that it
makes the solution sound trivial: get a better 'fi lter', and the problem
goes away. The diffi culty, of course, is that no fi lter yet exists that is
sophisticated enough to cope with the complexities of life science data
and the way in which it is described in 'human-readable' form. The
simple key word searches or faceted browsing techniques that work so
well for buying consumer goods or for booking a holiday online, break
down quickly when faced with complex, context-sensitive, inter-related,
incomplete and often ambiguous scientifi c data and articles. In many
Search WWH ::
Custom Search