Deep learning and health care

Into the deep world of medical data — ft. a brief history

Over the past decade, key terms such as artificial intelligence (AI), machine learning, neural networks, and deep learning have found their place in our daily vocabulary. However, while closely related, there are subtle yet significant differences between these terms. Hence, to precisely understand how data is ingested, analyzed, and returned, we must understand the difference between these terms. In this article, we will focus on modern technology’s oceanic term: deep learning. Then, considering modern medicine’s inevitable reliance on computing, we will focus on the applications of deep learning in health care.

What is Deep Learning?

Also known as deep structured learning or hierarchical learning, deep learning is a subfield of machine learning that uses a layered algorithmic architecture to analyze data. Since the algorithm is usually comprised of a large of number of layers (or neurons), it is referred to as deep.

• In deep learning models, data is filtered through a cascade of multiple layers, with each successive layer using the output from the previous one to inform its results. Deep learning models can become more and more accurate as they process more data, essentially learning from previous results to refine their ability to make connections and correlations.

Deep learning is loosely based on the way biological neurons connect with one another to process information in our brains. Similar to the way electrical signals travel across the cells of living creates, each subsequent layer of nodes is activated when it receives stimuli from its neighboring neurons.

Into the Deep Past — a brief history of deep learning

1943Logician Walter Pitts and neurophysiologist Warren McCulloch used the neural networks of the human brain to create a computer model based on it.

• Pitts and Warren employed a combination of mathematical equations and algorithms they called “threshold logic” to mimic the human thought process.

1960Mathematician Henry J. Kelley developed the basics of a continuous Back Propagation Model, introducing the concept of back propagation: the backward propagation of errors for training purposes.

1965The earliest developments of deep learning algorithms appeared.

• Ukrainian mathematician Alexey Grigoryevich Ivakhnenko developed the Group Method of Data Handling.
• Valentin Grigorʹevich Lapa wrote Cybernetics and Forecasting Techniques.
• They both used models with polynomial activation functions that were then analyzed statistically. From each layer, the best statistically chosen features were forwarded on to the next layer in a slow, manual process.

From the 1970s to the early 2000s, the world of machine learning experienced a widespread rejection from the scientific community as a result of its great promises (such as creating machines that operate as good as — or even better — than the human brain) that couldn’t be kept. As a consequence of the subsequent lack of funding, the development of both deep learning and artificial intelligence became limited. Despite this, there were still individuals who carried on the research without funding.

2001A research report by META Group (now called Gartner) described the increasing volume of data and the increasing speed of data generation. This was a call to prepare for handling these large volumes through innovations like deep learning.

2012 Google Brain released the results of The Cat Experiment. The use of deep learning in this project proved this tool’s significant advantages in terms of efficiency and speed. As a result, many other projects began implementing deep learning more commonly.

Currently, deep Learning is still evolving…

Deep Learning in Medicine

Over the past couple of years, healthcare has shown to immensely benefit from deep learning because of the large volume of data being generated (about 150 exabytes, or 1018 bytes, in the United States alone). However, deep learning can especially assist health care with medical diagnosis, and more specifically with medical imaging.

• Diagnostic mistakes are common. In fact, between 12 to 18 million Americans face some type of misdiagnosis each year. There is hope that deep learning can change this unsettling situation for the better.
• Medical imaging: In recent years, deep learning has been used to analyze medical images in various fields, showing excellent performance. One of the most widely used deep learning algorithms in this field is convolution, from which the convolutional neural network (CNN) is derived. This system is inspired by the primary visual cortex in animals. It has the capacity to decipher and learn the most complex patterns existing in a set of images, with applications in analyzing X-rays, MRIs, and CT scans. For example, in 2017, Stanford University computer scientists created a CNN model trained on 130,000 clinical images of skin pathologies to detect cancer.

With current advancements in the field of machine learning — and specifically deep learning — the field of health care seems to be able to take advantage of powerful tools. With this in mind, it is safe to say that the ever-evolving intersection of medicine and technology is undergoing a big change: a deep revolution!

Check out MIT Introduction to Deep Learning to learn more about the topic.

About That Medic Network

If you like this article, be sure to follow us for more content on our Website, Facebook, Twitter, Instagram, LinkedIn.

Feeling friendly or excited for our next article? Subscribe to our Medium publication to stayed tuned.

Email us at info@thatmedicnetwork.com if you have any questions.

References

1. Bresnick, Jennifer. “What Is Deep Learning and How Will It Change Healthcare?” HealthITAnalytics, HealthITAnalytics, 18 Dec. 2019, https://healthitanalytics.com/features/what-is-deep-learning-and-how-will-it-change-healthcare.
2. “Deep Learning in Medical Diagnosis: How Ai Saves Lives and Cuts Treatment Costs.” AltexSoft, AltexSoft, 23 Apr. 2020, https://www.altexsoft.com/blog/deep-learning-medical-diagnosis/.
3. Esteva, Andre, et al. “A Guide to Deep Learning in Healthcare.” Nature News, Nature Publishing Group, 7 Jan. 2019, https://www.nature.com/articles/s41591-018-0316-z.
4. Foote, Keith D. “A Brief History of Deep Learning.” DATAVERSITY, 31 Jan. 2017, https://www.dataversity.net/brief-history-deep-learning/.
5. Markoff, John. “How Many Computers to Identify a Cat? 16,000.” The New York Times, The New York Times, 25 June 2012, https://www.nytimes.com/2012/06/26/technology/in-a-big-network-of-computers-evidence-of-machine-learning.html.