Can AI lead the future of Medicine?
Technologies
are becoming ever-present in modern business and everyday life, giving birth to
modern structures like Amazon Go which requires no employee, and yet customers
can shop without any hassle. And this influence is not solely limited to such
business sectors, as with time the Healthcare field is also slowly being
modernized with Artificial Intelligence (AI) or also known as Deep Medicine.
The ability of a computer algorithm to come to a conclusion based on some input
data or training data is called Artificial Intelligence.
And,
when such machine learning algorithms are used to mimic human behavior,
condition, and other factors as input and deliver an output like diagnosis,
treatment protocol, etc. we call it Deep Medicine.
AI
has the potential to assist health providers in many ways and some even believe
that soon AI will replace human physicians. But such a perspective is still far
away in the future, as Deep medicines have strong relevance in the healthcare field
but the support they provide and the algorithms they use can vary significantly.
And
so, something like replacing doctors will still need time as there is a board
range of medical tasks that need to be handled to replace them. And, till now
no such AI has proofed to be able to do so with great efficacy.
So,
up until now, they have been only used to develop treatment procedures, drugs,
and patient monitoring, with physicians holding the final call to action, as it
is quite susceptible to errors.
But
that might soon be treated as the past! The recent evolution of IBM’s AutoML to
AutoAI might just be able to bring new innovation in the fields of Deep
medicine.
What is IBM AutoAI?
Automated
Artificial Intelligence (AutoAI) is a variation of the automated machine learning or AutoML,
technology, which extends the automation of model building towards automation
of the full life cycle of a machine learning model.
It
applies intelligent automation to the task of building predictive machine learning models by
preparing data for training, identifying the best type of model for the given
data, then choosing the features, or columns of data, that best support the problem the model is solving.
Finally,
automation tests a variety of tuning options to reach the best result as it
generates, then ranks, model-candidate pipelines. The best-performing pipelines
can be put into production to process new data, and deliver predictions based
on the model training. It can also be told to make sure that there is no
inherent bias and constantly improve the model.
This machine-learning system complements the previous Watson studio, helping it to
simplify an AI lifecycle management, AutoAI automates:
- ·
Data preparation
- ·
Model development
- ·
Feature engineering
- ·
Hyper-parameter optimization
And
produces pipelines of potential predictive systems. Using those pipelines data
can be calculated at very high accuracy to obtain the desired value.
Future Possibilities of AI in the fields of Healthcare -
The ability of AutoAI to develop a predictive machine learning model has quite the
potential to develop many fascinating medical applications and systems. An
example of such an application maybe –
- · An app to predict the possibility of a person having a heart attack or heart failure.
- ·
Predict the health insurance cost etc.
According
to WHO Every year an estimated 17.9 million lives are lost to Cardio-vascular Disease
(CVD) alone, making it the number one cause of global death. And in the coming
years, it is sure to surge as the lifestyle that we opt to maintain is the root
cause of such high cases.
CVDs
are comprised of a group of disorders related to the heart and blood vessels. And
among them, Heart attack takes the spot as the deadliest, as 4 out of 5 CVDs
deaths are due to this lurking danger.
But
problems like heart attack and heart failure don’t occur all of a sudden. They
are progressive disease. Their roots start to develop a long time before
their effects are seen. And, multiple unaccounted factors influence them, some
even from as early as childhood like unhealthy diet and obesity, physical
inactivity, etc.
And
again, it isn’t possible for all of us to always have a Doctor check on us for
any ill health development and we, the general people, can’t also diagnose
ourselves ahead of time. As such the problems like CVDs continue to persist in
our society without any signs of backing down.
But
if technologies like AutoAI are trained using data of CVD patients then the
predictive models will be able to tell whether a person has the possibility of
developing any heart disease or not.
And
the best part is such models can be converted into easy-to-use mobile or
computer applications using simple Java scripts or other codes using IBM itself.
Which can later be used by general people for a gross estimation of their
health status. Improving early diagnosis of disease and overall health of the
community.
With the advancement of science and technology, we have found solutions to many
problems which seemed impossible in the past. But as we progress new obstacles
keep on growing on our path. And such is also the case for our healthcare
sector.
And
thus, for again creating a revolution in our health sector new innovative
methods are needed. And one such may be the AutoAI of IBM.
Although
still hidden under shades of doubt and with obvious flaws like compromised data
privacy, automation of jobs, lack of adequate authentic data for training. This
technology still shows us a glimmer of hope.
It
is hoped that one-day such technology can provide us with insight into factors
which we normally don’t consider and from a very early stage too and help
develop personalized treatment procedure according to the condition of the
patients, preventing undue exposure to chemical products and therapies.
Reference –
- 1. https://www.who.int/health-topics/cardiovascular-diseases/#tab=tab_1
- 2. https://bmcmedinformdecismak.biomedcentral.com/articles/10.1186/s12911-020-1023-5
- 3. https://dataplatform.cloud.ibm.com/docs/content/wsj/analyze-data/autoai-overview.html
- 4. https://en.wikipedia.org/wiki/Artificial_intelligence_in_healthcare
Software Used –
1. IBM
Watson Studio AutoAI - https://www.ibm.com/demos/collection/IBM-Watson-Studio-AutoAI/?lc=en
0 Comments