QUICK FACTS
Created Jan 0001
Status Verified Sarcastic
Type Existential Dread
Keywords
data engineering, engineering, knowledge, stop sign, electromagnetism, machine learning, artificial intelligence, control theory, signal processing, microelectronics

Information Engineering

“Right. Let's get this over with. Don't confuse this with Data engineering; that's a different brand of organized...”

Contents
  • 1. Overview
  • 2. Etymology
  • 3. Cultural Impact

Right. Let’s get this over with. Don’t confuse this with Data engineering ; that’s a different brand of organized chaos.

Information engineering is the engineering discipline that concerns itself with the generation, distribution, analysis, and—most optimistically—the use of information, data, and knowledge within electrical systems. It’s a field that crawled into existence in the early 21st century, mostly by slapping a new, ambitious label on a collection of problems people were already failing to solve.

Here’s a computer attempting to identify a stop sign . A monumental achievement in telling a machine what you should have learned in driver’s ed.

The sprawling umbrella of information engineering covers a litany of fields. On the more theoretical, head-in-the-clouds end, you have disciplines like electromagnetism , the ever-fashionable machine learning , its overconfident cousin artificial intelligence , the Sisyphean task of control theory , the art of finding meaning in noise known as signal processing , and the microscopic tyranny of microelectronics . On the more applied, “let’s make it do something before the funding runs out” end, you’ll find computer vision , natural language processing , bioinformatics , medical image computing , cheminformatics , autonomous robotics , its less ambitious sibling mobile robotics , and the tangled web of telecommunications . A significant number of these fields were poached from computer engineering , with others borrowed from adjacent disciplines like electrical engineering , computer science , and bioengineering , because originality is overrated.

A visual representation of clustering , which is just a mathematical way of putting things in boxes. Humans do it with labels; machines do it with data points. The impulse is the same.

At its core, the field of information engineering is propped up by a formidable scaffold of mathematics. It leans, rather heavily, on the unforgiving logic of probability , statistics, calculus , linear algebra , optimization , differential equations , variational calculus , and complex analysis . It’s a testament to humanity’s willingness to throw increasingly complicated math at a problem in the hopes that it will eventually surrender.

Information engineers, the people who willingly subject themselves to this, often possess a degree in information engineering or a closely related field of study. They frequently join a professional body like the Institution of Engineering and Technology or the Institute of Measurement and Control , presumably for the professional validation and the newsletters. A citation is needed for how often this happens, but let’s be honest, the desire for belonging is a powerful motivator. Given the pervasive nature of information engineering, these professionals are found in nearly every industry, a quiet testament to how widespread the need to manage data has become.

History

During the quaint decades of the 1980s and 1990s, the term “information engineering” was used to describe a specific domain within software engineering. By the 2010s and 2020s, that same domain had been rebranded into what is now known as data engineering . It’s the same wine, just in a new, trendier bottle.

Elements

Machine learning and statistics

Machine learning is the discipline dedicated to using statistical and probabilistic methods to enable computers to “learn” from data without someone having to painstakingly program every single rule. It’s less about genuine learning and more about sophisticated pattern matching. The related field of data science involves applying these machine learning techniques to dredge knowledge from vast oceans of data, often with the hope of finding something, anything, useful.

The subfields are a veritable alphabet soup of approaches, including deep learning , where the models are as inscrutable as they are powerful; supervised learning , which is basically learning with an answer key; unsupervised learning , which is the equivalent of telling a machine to “figure it out”; reinforcement learning , a trial-and-error process of digital rewards and punishments; semi-supervised learning , for when you can’t be bothered to label all your data; and active learning , where the machine gets to ask questions, which is just as annoying as it sounds.

Causal inference , the dark art of distinguishing correlation from causation, is another critical, and often overlooked, component of this domain.

Control theory

Control theory is fundamentally about the control of (continuous ) dynamical systems . The entire goal is to prevent things from going wrong—to avoid delays, overshoots, or the catastrophic loss of stability . Information engineers typically fixate on the theoretical underpinnings of control, leaving the messy, physical design of control systems and circuits to the electrical engineers. It’s cleaner that way.

Its subfields include classical control , with its time-tested methods; optimal control , which seeks the “best” possible solution under a set of constraints; and nonlinear control , for when the universe refuses to behave in a simple, linear fashion.

Signal processing

Signal processing deals with the generation, analysis, and application of signals . These signals can manifest in countless forms, whether as an image , a sound , an electrical impulse, or a biological marker. It is the science of extracting clarity from a world of static.

A demonstration of how the 2D Fourier transform can be applied to surgically remove unwanted information from an X-ray scan . It’s like Photoshop for diagnosticians.

Information theory

Information theory is the rigorous study of the analysis, transmission, and storage of information. It’s obsessed with quantifying information and figuring out the absolute limits of how efficiently it can be handled. Its major subfields are coding —the creation of rules for representing information—and data compression , the endless quest to say the same thing with fewer bits.

Computer vision

Computer vision is the ongoing, often frustrating, attempt to make computers understand image and video data at a high, almost human-like, level. It involves teaching a machine the difference between a cat and a coffee mug, a task that has proven to be profoundly more difficult than anyone initially imagined.

Natural language processing

Natural language processing is concerned with getting computers to comprehend human (natural) languages. This endeavor is complicated by the fact that human language is a messy, context-dependent, and often illogical construct. It usually involves processing text , but frequently extends into the domains of speech processing and recognition , adding another layer of acoustic ambiguity to the problem.

Bioinformatics

Bioinformatics applies the tools of information engineering to the analysis, processing, and use of biological data. It’s the field where computer science collides with the chaotic, wet code of life itself. This typically involves complex topics like genomics and proteomics , and sometimes bleeds into the territory of medical image computing .

Cheminformatics

Cheminformatics is the less glamorous but equally important sibling of bioinformatics. It focuses on the analysis, processing, and use of chemical data, applying computational methods to understand molecules and their interactions.

Robotics

Within the context of information engineering, robotics is less about building physical machines and more about programming the minds that control them. The primary focus is on the algorithms and computer programs that give robots their autonomy. Consequently, information engineering tends to gravitate towards autonomous, mobile, or probabilistic robots, which present the most interesting computational challenges. Major subfields that occupy information engineers include control , perception (seeing the world), SLAM (building a map while simultaneously trying not to get lost in it), and motion planning (figuring out how to get from A to B without crashing into things).

Tools

Once upon a time, some corners of information engineering, like signal processing, relied on the charmingly archaic world of analog electronics . Today, that’s a historical footnote. The vast majority of information engineering is executed on digital computers . Many of the tasks are embarrassingly parallelizable , which has led to a hardware arms race. Modern work is carried out on a combination of CPUs , GPUs , and a burgeoning market of specialized AI accelerators . More recently, there has been a speculative, almost sci-fi interest in leveraging quantum computers for certain subfields, such as quantum machine learning and quantum robotics , though their practical application remains, for now, largely theoretical.

See also