Learning and Triage for the Health Internet of Digital Twins

Dear Colleagues,

The medical sciences involve natural signals, without an immediate and exact relation to the ingredients of a mathematical model. The perfect algorithm can be too complex for real-time execution, while the fast solution lacks the required accuracy. Spreading the algorithmic ingredients in time and space results in many architectural arrangements that are worthy of consideration. For example, it has been demonstrated that the blood pressure meter can be bettered, from 25% for the single device to 3%. This uses a judicious selection of many crude measurements, each sharing packages with other parts. This is the world of the Internet of Things.

The human body is the most famous example of a natural IoT, or, in other words, “Internet of Organs”. Electronic sensors can be used at a number of body locations but will deliver changing results. There is general agreement on the measurement procedure, but little effort has been made to normalize the results. Blood pressure is a commonly quoted health value measured at the left wrist, but the missing location information results in 8% deviation.

One suitable engineering approach comes from the Industry 4.0 philosophy of “Digital Twins”. The digital model is developed hand-in-hand with reality, carefully checking each enhancement with the original till the model has sufficient detail. The strength of this approach is that software mock-ups are realistic.

Artificial neural learning uses a Hebb-like update in conjunction with a measure of arithmetic error. This concept cannot be readably extended to the Internet of Things. This can be repaired by (1) linking the network and the formal accuracy model, and (2) extending the update mechanism to two levels. A related benefit is the crystal-clear handling of the difference between neural and statistical data processing.

There are more mathematical issues at play. A thorough understanding of numbering systems reaches deep into the system behavior. This is demonstrated especially well in Chua’s memristor. The concept should be extended towards higher levels of abstraction, beyond cell arrays. Similar issues are found in computational techniques.

Learning is a kind of copying. Infant learning is fed by memorization, which is reproduced or even adapted until understanding can be shown by generating new material. The importance of closing the gap between student and master lies in being able to renew learning in a changing environment.

Another reason for continuous learning is the imperfection of sensing. There is much variance based on design and fabrication. Under many circumstances, complex signal processing is required to extract the target signal and/or determine a desired signal set. Failures in synchronization encourage successful learning.

What is good for the goose, should also be good for the gander. What can be learnt from one technology to make improvements in another? What is needed to create artificial awareness in an E-bicycle and would this reduce the number of accidents? Additionally, what are the technical challenges that must be overcome? What is the biological equivalent of the GALS principle? Is the Digital Twin of the human body feasible? We invite contributions to this Special Issue that examine the multi-disciplinary phenomena that facilitate triage devices in human health. An example is the role of reference examples. For instance, when you have a constant body temperature, a higher temperature may be alarming. You do not always need the exact value: that can wait!

Learning by Reference takes physical plausibility into account for reading a measurement with increased accuracy. For instance, when compared with walking in the sun, the skin temperature will be lower in the shade. The skin temperature for a wrist sensor has a similar deviance. However, the offset is not a constant in a body network, as feature synchronization has to be included. This Special Issue will include such Learning by Reference mechanisms, including, but not limited to, swarming for triage in a typical polyclinic.

Prof. Dr. Lambert Spaanenburg
Guest Editor

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Electronics is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.