Successful machine learning (ML) applications depend on hand-engineering features where the researcher manually encodes relevant information about the task at hand and then there is learning on top of that.
The deep learning (DL) researcher will try and get the system to engineer its own features as much as is feasible. Unlike machine learning, deep learning is mostly unsupervised. It involves, for example, creating large-scale neural nets that allow the computer to learn and “think” by itself without the need for direct human intervention.
Which one is the essential one in Health 4.0, that means in medicine discovery, prevention and participative medicine depends on the evolution of technology and particularly on the training of health workers on those essential aspects, allowing its implementation and ethics analysis.
The H2O, humanization of healthcare is based at the moment in Machine Learning, particularly because is mainly focused on circulatory diseases and there is a lot scientific work demonstrating risk factors in the field that can be obtain from non-intrusive sensors
Nevertheless most of the experiences of AI tools that up to now have been applied to Medicine and patient care were based in decision trees and bayesian statistics based on Evidence Based Medicine (EBM) and never on the premises to allow Deep Learning on individual and personal health data throughput (pHSD= personal Health Small Data), in which privacy and liability are essential.
The situation at the moment is not very promissing since Health care workers do not study those principles in their carriers and are not trained on QoC (quality of care) premises based on IoT (Internet of the things). Nevertheless there is always innovation groups that try to fix the most advance principles for patient´s empowering.
Inspired in: Why Google is investing in Deep Learning.
