Using Artificial Intelligence / Machine Learning in Medicine

Shubham Mahendra Aditya Dipesh Prasad Sanveer Katriyar

Wakpaijan [email protected] sanveerkatriyar46@gmail
shubhamwakpaijan@gmail. .com
com
B.Tech, Department of Computer Science and Engineering, SOET
D Y Patil University, Ambi

Abstract: 1. Introduction
This research paper explores the use of
Artificial Intelligence (AI) and Machine Artificial intelligence (AI) in varying forms and
Learning (ML) techniques in the field of degrees has been used to develop and advance
advanced medicine. The paper highlights the
a wide spectrum of fields, such as banking and
potential of AI/ML in revolutionizing medical
practices, improving patient outcomes, and financial markets, education, supply chains,
enhancing healthcare delivery. It examines manufacturing, retail and e-commerce, and
various applications of AI/ML, including
healthcare. Within the technology industry, AI
diagnosis, treatment, drug discovery, medical
imaging, and patient monitoring. The has been an important enabler for many new
challenges and ethical considerations business innovations. These include web search
associated with AI/ML in advanced medicine
(e.g., Google), content recommendations (e.g.,
are also discussed. Through a comprehensive
review of existing literature and case studies, Netflix), product recommendations (e.g.,
this paper aims to provide insights into the Amazon), targeted advertising (e.g., Facebook),
current state and future directions of AI/ML in and autonomous vehicles (e.g., Tesla).
advanced medicine.

Keywords: Humans reap the benefits of artificially

Artificial Intelligence, Machine Learning, intelligent systems every day. Starting from the
advanced medicine, diagnosis, treatment, drug spam free emails that we receive in our
discovery, medical imaging, patient
inboxes, to smart watches that use inputs from
monitoring.
accelerometer sensors to distinguish between
mundane activities and aerobic activity, to
buying products on online shopping sites, like
Amazon that recommend products based on
our previous purchase records. These examples
represent the use of AI in a variety of fields,
such as technology and retail. AI has bot assistant that can translate languages,
transformed our everyday lives, with an effect transcribe notes, and organize images and files.
on the way we perceive and process
information.

This article aims to present various aspects of AI

as it pertains to the medical sciences. The
article will focus on past and present day
applications in the medical sciences

3. Present Day Use of AI/ML:

2. What is AI ?

The most recent application of AI in global

AI is a wide-ranging branch of computer science
healthcare is the prediction of emerging
concerned with building smart machines
hotspots using contact tracing, and flight
capable of performing tasks that typically
traveller data to fight off the novel coronavirus
require human intelligence. Some applications
(COVID-19) pandemic.
of AI include automated interfaces for visual
perception, speech recognition, decision-
making, and translation between languages. AI
is an interdisciplinary science.

Artificially intelligent systems in healthcare have

the following typical pattern. Such a system
starts with a large amount of data, on these
data machine-learning algorithms are employed
to gain information, this information is then Contact tracing is a disease control measure
used to generate a useful output to solve a used by government authorities to limit spread
well-defined problem in the medical system. of a disease. Contact tracing works by
It captures the typical workflow of an AI contacting and informing individuals that have
solution. Applications of AI in the field of been exposed to a person who has contracted
medical sciences include matching patient the disease and instructing them to quarantine
symptoms to appropriate physician, patient to prevent further spread of the disease. As
diagnosis, patient prognosis, drug discovery, reported by Apple Newsroom, tech giants like
Google and Apple have joined forces to create a
contact tracing platform that will use artificial quickly process large amounts of data and
intelligent systems through the use of identify patterns or abnormalities that might be
application programming interfaces commonly missed by human observers. This helps doctors
referred to as API's on smartphones. The make more accurate and timely diagnoses,
platform will enable users who choose to enrol leading to better patient outcomes.
to report their lab results. Location services will
then allow the platform to contact people who
Predictive Analytics: AI/ML can predict patient
may have been in the vicinity of the infected
outcomes by analysing a vast array of data,
person.
including electronic health records (EHRs),
genetic information, and lifestyle factors. By
4. Future of AI/ML in the field of identifying risk factors and patterns in the data,
AI/ML models can help doctors predict the
Medicine:
likelihood of diseases, complications, or adverse
Artificial intelligence (AI) and machine learning events. This enables proactive interventions and
(ML) have revolutionized many industries, personalized treatment plans.
including medicine. In healthcare, AI and ML
technologies are being increasingly utilized to
improve diagnostic accuracy, enhance Drug Discovery and Development: AI/ML
treatment planning, optimize resource techniques are transforming the drug discovery
management, and advance patient care. Here's process, making it faster and more efficient. AI
an introduction to the use of AI/ML in medicine: algorithms can analyse vast databases of
molecular structures and identify potential drug
candidates with higher success rates. ML
models can also predict the effectiveness of
new drug compounds, reducing the need for
extensive laboratory testing. This accelerates
the development of new treatments and
therapies.

Diagnostic Assistance: AI and ML algorithms Personalized Medicine: AI/ML enables

can assist healthcare professionals in diagnosing personalized treatment plans based on
diseases by analysing medical images, such as X- individual patient characteristics. By analysing
rays, CT scans, and MRIs. These algorithms can
patient data, including genetics, medical history,
and treatment responses, AI/ML algorithms can
It's important to note that while AI/ML
identify optimal treatment options tailored to
technologies hold great promise, they are not
each patient. This approach maximizes
meant to replace healthcare professionals.
treatment effectiveness while minimizing
These technologies are designed to augment
adverse reactions, providing more precise and
human expertise and improve decision-making
patient-centric care.
processes, ultimately leading to better patient
outcomes and more efficient healthcare
delivery.
Remote Monitoring and Telemedicine: AI/ML
technologies facilitate remote patient
monitoring and telemedicine. Wearable devices
and sensors collect real-time patient data,
which is then analysed by AI algorithms. This
5. Myth v/s Reality:
allows healthcare providers to remotely
monitor patients' vital signs, detect anomalies,
and intervene in case of emergencies.
Telemedicine platforms powered by AI can
provide virtual consultations, enabling patients
to access healthcare services from anywhere,
Myth 1: AI will replace doctors and healthcare
increasing accessibility and reducing healthcare
professionals.
costs.
Reality: AI is not intended to replace doctors or
healthcare professionals. Instead, it is designed
Healthcare Administration and Workflow to assist and augment their capabilities. AI
Optimization: AI/ML can streamline systems can analyse medical data, provide
administrative tasks, such as medical coding and insights, and support decision-making, but the
billing, appointment scheduling, and resource final treatment decisions and patient care are
allocation. By automating these processes, still made by healthcare professionals. AI serves
healthcare organizations can reduce errors, as a valuable tool to enhance their expertise
enhance efficiency, and allocate resources more and improve patient outcomes.
effectively. This frees up healthcare
Myth 2: AI will lead to job losses in the
professionals' time, allowing them to focus on
healthcare industry.
patient care.
Reality: While AI may automate certain tasks, between researchers, healthcare providers, and
its primary aim is to augment healthcare policymakers to ensure responsible and
professionals, not replace them. AI technologies effective implementation.
can alleviate the administrative burden,
enhance efficiency, and enable healthcare
providers to focus more on patient care. Instead
6. Limits and Challenges:
of job losses, AI is expected to reshape job roles The application of artificially intelligent systems
and create new opportunities for healthcare in any field including healthcare comes with its
professionals with specialized AI knowledge. share of limitations and challenges. The time
has come to change our mindset from being
reactive to being proactive with regards to
Myth 3: AI algorithms are biased and
downfalls of new technology. Here we discuss
discriminatory.
those challenges focusing more on those that
Reality: AI algorithms can inherit biases from
pertain particularly to healthcare.
the data they are trained on. It is crucial to
address this issue by carefully curating diverse
and representative training data and Availability of data :

implementing fairness measures. Researchers The first step towards building an artificially
and developers are actively working to reduce intelligent system (after problem selection and
biases in AI algorithms and ensure equitable development of solutions strategy) is data
healthcare outcomes. Ethical considerations and collection. The creation of well performing
ongoing monitoring are essential to mitigate models relies on the availability of large
potential biases in AI systems. quantities of high-quality data. The issue of data
collection is shrouded in controversy due to
patient privacy and due to recent incidents of
Myth 4: AI will completely revolutionize
data breaches by major corporations. Advances
healthcare overnight.
in technology have resulted in increased
Reality: While AI has significant potential in
computational and analytic power as well as the
healthcare, its full integration and impact will
ability to store vast amounts of data.
take time. Implementing AI technologies
requires addressing technical, ethical, and
regulatory challenges. The adoption and Creating biased models-Biased data :

integration of AI in healthcare will be a gradual Artificially intelligent systems are then trained
process, evolving through collaborations with a portion of the data that was collected
(also known as training data set) with the Model selection :
remaining data reserved for testing (also known With the existence of several algorithms and
as testing data set). Thus, if the data collected is models to choose from, one must select the
biased, that is, it targets a particular race, a algorithm that is best suited for the task at
particular gender, a specific age group then the hand. Thus, the process of model selection is
resulting model will be biased. Thus, the data extremely important. Bias models are ones that
collected must be a true representation of the are overly simple and fail to capture the trends
population for which its use is intended. present in the dataset.

Presenting biased models

It is important for a user of an artificially

intelligent system to have a basic understanding
of how such models are built. This way a user
can better interpret the output of the model
and decide how to make use of the output. For
instance, there are many metrics that one could
use to evaluate the performance of a model,
such as accuracy, precision, recall, F1 score,
Data pre-processing
and AUC score.[21] However, not every metric is
Even after unbiased data has been collected, it
appropriate for every problem. When the user
is still possible to create a biased model. The
of an artificially intelligent system is presented
collected data must be preprocessed before it
with performance metrics of a model, they
can be used to train an algorithm. The raw data
need to make sure that the metrics appropriate
that has been collected often contains errors
to the problem are being presented and not just
due to manual entry of data or a variety of
the metrics with the highest scores.
other reasons. These entries are sometimes
modified through mathematical justification or
are simply removed. Care should be taken that Fragmented data

data preprocessing does not result in a biased Another limitation of the application of AI is
pool of data. that models that one organization spends time
and effort to design and deploy for a specific
task (regression, classification, clustering, NLP,
etc) cannot be seamlessly transitioned for
immediate use to another organization without
recalibration. Due to privacy concerns, data equitable use of AI in medicine. Collaboration
sharing is often inaccessible or limited between between healthcare professionals, researchers,
healthcare organizations resulting in policymakers, and technology developers will be
fragmented data limiting the reliability of a crucial for the successful integration of AI in
model. medical sciences.

While AI is a powerful tool, it is not intended to

Black boxes : replace healthcare professionals. Human

expertise, critical thinking, and compassion
Artificial Intelligent systems have a reputation of
remain vital in healthcare decision-making and
being black boxes due to the complexity of the
patient care. The future of AI in medical
mathematical algorithms involved. There is a
sciences lies in the collaboration between AI
need to make models more accessible and
technologies and human expertise, ultimately
interpretable. While there is some recent work
leading to more precise diagnoses, personalized
in this direction, there is still some progress to
treatments, improved patient care, and
be made.
advancements in medical research.

CONCLUSION: The Future of AI/ML

REFERENCES:
in Medical Science
1. Geéron A. Hands-On Machine Learning
In conclusion, the future of AI in medical
with Scikit-Learn & Tensor Flow. 1st ed.
sciences is highly promising. AI technologies and
Sebastopol, CA, USA: O'Reilly; 2017.
machine learning algorithms are revolutionizing
2. McCarthy J, Minsky M, Rochester N,
various aspects of healthcare, including
Shannon C. A proposal for the
diagnosis, treatment, drug discovery, patient
Dartmouth summer research project on
monitoring, and healthcare administration. By
artificial intelligence 195
leveraging large-scale data analysis, predictive
3. Ekins S. The next era: Deep learning in
analytics, and pattern recognition, AI can
pharmaceutical research. Pharm
enhance precision medicine, optimize
Res. 2016
treatment plans, and improve patient
outcomes.

However, it is essential to address ethical

considerations, data privacy, and regulatory
frameworks to ensure the responsible and