New Insights:
Understanding the Nature of Information
Some of the most important paradigms in Physics and in Information Sciences over the last century have just been turned on their head in a manuscript released. In this paper, "Information: to Harvest, to Have and to Hold", Marin van Heel (CNPEM, Campinas, Brazil) and Michael Schatz (Image Science), have re-defined the classical concepts of the Signal-to-Noise Ratio (SNR) and Claude Shannon's concept of Information capacity. The authors have replaced those metrics by realistic, measurable metrics. Applying the new metrics, they were capable of visualising the glycan-shield of the SARS-CoV-2 spike proteins, which cloak the virus uses to stay hidden from the human immune system. Without adding any new biochemical research of their own, they use concepts like the "Local Information Density" to compare the research results that others have recently deposited in public data banks.
|
Marin van Heel and Michael Schatz have shown that the SNR, a metric that was seen
as almost synonymous to Information by scientists of all trades for over a century,
can only be used to roughly estimate how much of the information we had at the entrance
of the channel has been lost in transfer. The SNR cannot be used to assess new
information one has obtained by, say, by taking X-ray pictures of patient's chest.
Claude Shannon's famous "channel information capacity" theorem, that is based on the
SNR, is thus also about not losing information in the process of sending that information
through a noisy channel like a telephone line or writing and retrieving data to a computer
hard drive.
The harvesting of information, (electron) microscopy, medical detection like X-ray or NMR |
diagnostics, astronomy imaging, cameras etc.) is different in nature this information is in
the first a signal-processing metric: it follows a modern perspective of physics where
the probability of detecting arriving electrons or photons is given by the square of
the wave function impinging upon the detector.
This new information perspective may change many aspects of modern science. It changes the way resolution is in data defined and how it is appreciated. The resolution levels achieved and the Local Information Density are different in different areas of the object. The new metrics may change the way we optimise medical X-ray instrumentation in order to minimise patient dose while collecting sufficient information for a specific diagnostic purpose, etc. |