AI for Immunology

Louis J. Catania

• Publisher: CRC Press
• Year: 2021
• eBook ISBN: 9781000369915
• Topic: Ciencia de la computación
• Subtopic: Redes neuronales

The bioscience of immunology has given us a better understanding of human health and disease. Artificial intelligence (AI) has elevated that understanding and its applications in immunology to new levels. Together, AI for immunology is an advancing horizon in health care, disease diagnosis, and prevention. From the simple cold to the most advanced autoimmune disorders and now pandemics, AI for immunology is unlocking the causes and cures.

Key features:

• A highly accessible and wide-ranging short introduction to AI for immunology
• Includes a chapter on COVID-19 and pandemics
• Includes scientific and clinical considerations, as well as immune and autoimmune diseases

1

Understanding Artificial Intelligence (AI)

Introduction

The Merriam-Webster dictionary defines AI as “…a branch of computer science dealing with the simulation of intelligent behavior in computers; the capability of a machine to imitate intelligent human behavior.”¹ Those simple words bespeak the magnitude of the science of AI. For “…a machine to imitate intelligent human behavior” is to have it effectively “mimic” the functions of the human brain or restated, in biologic terms, to mimic neuroscience. The neurological functions associated with intelligence are related to the cortical control centers, progressive neural layers, and the complex neural networking between these centers and layers. Thus, AI must simulate the structures of these cortical centers and layers and their neural network functioning.

While there are substantial differences between AI and basic computers, there also are some common denominators in their structure and function. The most relevant among them include three fundamental categories: (1) the input layer, components for entering “user data” (e.g., hardware such as keyboard, printers, audio/visual); (2) the inner (hidden) layer, a functional component for data processing (sequence of automated operations to convert user input data into computer language or “code”) consisting of hardware like microprocessors (central processing unit [CPU], random access memory drive [RAM]) and software (“hidden” operating systems [OS] programming applications); and (3) the output layer consisting of the hardware devices (monitors, printers, audio/visual) to present computed results.

A Model for Understanding Artificial Intelligence

AI utilizes multiple hardware and software tools in its inner layer to effectively mimic the neurological functions associated with intelligence. These inner layer functions are classified into the two main categories of AI: “machine learning” and “deep learning.” These two categories introduce unique aspects to the computing process that electronically reproduce the qualities of human intelligence. To accomplish this task, AI uses “algorithms,” mathematical formulas to simulate the progressive layers of neuronal functions and neural networking in the human brain.

A simple way (kind of) to understand how AI works is to briefly present how the structures and functions of the human brain process information. Then we can compare that process to the structures and functions associated with the three layers of AI computing and its machine and deep learning process.

The fundamental neuroanatomical component of the brain that dictates neural functioning is called “the neuron.” It is estimated that there are approximately 100 billion neurons in the human brain.² Nerve impulses travel down axons reaching junctions called synapses where neurotransmitter chemicals are released across the synaptic cleft activating other neurons. All of this activity can be reduced to a mathematical model using linear algebra and differential calculus³ (Figure 1.1).

Figure 1.1 Mathematical model of neuron. (Diaa, Ahmedien [2018] The Mathematical Model of the Biological Neuron.)

The brain receives its user (input) data which it transmits through inner layer neuronal pathways in nuclei called the limbic system (analogous to a computer operating system [OS]). These nuclei act as relay stations that analyze input data and transmit it to appropriate higher cortical layers for interpretation. This vast network of 100 billion interconnecting neurons creates a complex in the human brain, collectively referred to as the neural network (NN), analogous to AI algorithms. This vast network produces multilayered, convolutional interconnections called the convolutional neural network (CNN).

The CNN has the potential of producing 100 trillion neural connections.⁴ This level of neural networking is called the “deep neural network” (DNN). The DNN is the analog of the AI electronic computing functions conducted by a central processing unit (CPU) and/or a graphic processing unit (GPU); programming applications; algorithms; databases; and servers. In the human brain, the results are human intelligence (output) while the electronic computer produces AI output. The sum total of this computer process is referred to broadly as “machine learning” with the specific CNN and DNN processes called “deep learning.”

These comparative analogies introduce hardware and software components associated with the science and technologies of AI. To complete an understanding of AI, descriptions of the major software and hardware components are needed. With that understanding, you will be able to appreciate the influences and applications AI has in the science of immunology.

AI Hardware

The guiding principle and goal of AI hardware technology are to support the enormous volume of data processing and the calculations and computations the AI software algorithms must execute simultaneously and in milliseconds. A brief thumbnail of each of the significant forms of current and evolving hardware will provide an understanding of the physical resources that drive the AI computing process and its diverse and complex software algorithms.

RAM (Random Access Memory)

A RAM microchip is a high-speed type of computer memory device that temporarily stores all input information as well as application software instructions the computer needs immediately and in the near future. RAM is read from any layer of the computer at almost the same speed. It loses all of its dynamic memory when the computer shuts down.⁵

When thinking of RAM relative to AI computing, consider it analogous to the hippocampus nucleus of the limbic system in the neuroscience model. It uses an ANN process called “memory networking” (similar to brain “plasticity”) to differentiate and adjust the connections between machine (unsupervised) and deep (supervised) learned information.⁶

Central Processing Unit (CPU)

The central processing unit (CPU) is the “brain” of the computer. It transforms the raw data into binary code (computer language) that can be manipulated and stored as memory in RAM. The CPU is the analog to the human brain at large when using our neuroscience analogy. It is the processor for all the operations the computer conducts. How your computer operates is based on mathematical operations (algorithms), and the CPU controls all of them through its arithmetic logic unit (ALU).

Graphic Processing Unit (GPU)

A GPU is a specialized microprocessor optimized for displaying graphics and doing very specific computer tasks. CPUs and GPUs are both made from hundreds of millions of transistors which can process thousands of operations per second. The GPU uses thousands of smaller and more efficient cores than the CPU and can handle multiple functions of lively parallel data at the same time. GPUs are 50–100 times faster in tasks that require multiple parallel processes, such as computer graphics and gaming (for which they were initially developed by Nvidia), but its most significant value is in its iterative computations of massive data load in machine learning, deep learning, and big data analytics.⁷

Servers

The term “server” simply means a machine (computer) that “serves” other machines, thus the name “server machine.”⁸ Whereas processors and CPUs are “the brain of the computer,” servers are their “heart.” They are computers themselves (hardware supporting software) designed to process requests and deliver data to another computer, a local network, or over the Internet.

Internet of Things (IoT)

A “hybrid” system known as the “Internet of Things” (IoT) is a system of interrelated computing devices, mechanical and digital machines, objects, and even people and animals that are provided with unique identifi...

Table of contents

• Cover
• Half Title
• Series Page
• Title Page
• Copyright Page
• Dedication
• Table of Contents
• Preface
• Acknowledgments
• Author
• 1 Understanding artificial intelligence (AI)
• 2 AI and the bioscience and clinical considerations for immunology
• 3 AI and chronic inflammation
• 4 AI and autoimmunity
• 5 AI and immunology considerations in pandemics and SARS-CoV-2 COVID-19
• 6 Emerging trends and future directions for AI in immunology
• Index