Noise hides in plain sight

Information is highly valued, therefore, it is no surprise why it gets a high financial and emotional price. There are different ways of getting access to information which range from buying books and textbooks, going to conferences, getting an education, paying a private tutor, paying for Internet access, being part of a particular culture, paying to read research articles etc. Although information is ubiquitous the quality and value are usually negotiated between the receiver and the sender. This, of course, does not mean that all information is available to everyone at all times. Although we can be ignorant of certain information, we can also be ignorant of being ignorant of that particular information as well. The possession of information does not necessarily imply well-informed since information can also, unknowingly to us, misinform us.

Cognitive science studies the brain as an information processing organ, whose unobservable functions can be scientifically studied via observable behaviors. This is in contrast to Behaviorism which attributes causality of behaviors to the environment, but does not theorize about brain processes, except to say that neurons change over time.¹ Cognitive scientists, with modifications, use concepts from the field of computer science in explaining, designing, and testing models of brain processes and human behavior. One of the most important concepts borrowed from the computer science field was Claude Shannon's information theory,²

Information theory studies the transmission, processing, extraction, and utilization of information. Abstractly, information can be thought of as the resolution of uncertainty. In the case of communication of information over a noisy channel, [...] "information" is thought of as a set of possible messages, where the goal is to send these messages over a noisy channel, and then to have the receiver reconstruct the message with low probability of error, in spite of the channel noise.³

Gallistel and King have studied how the brain, over noisy environments, extracts information and creates unobservable mental representations either for immediate or later use.⁴ If the aim of a system is to be accurate, understanding the complex relationship between signal to noise and bias is of crucial importance. This difficult to study relationship, which has been studied under decision theory, was illustrated by Daniel Kahneman et al. in a recent article.⁵

Kahneman et al. state that noise is "always undesirable—and sometimes disastrous", they add that it is beneficial "to know about bias and noise" in decision-making, but this can be challenging due to difficulties in error measurement and uncertainty of decision outcomes.⁵ If decision making is a function of our mental representations from acquired information, how our brains process information and the source of information need to be examined more closely as well. It is not uncommon to find medical information in different settings, including lectures and books, displayed without its evidence as follows,

Signs and symptoms for atrial fibrillation:

• Fatigue
• Tachypnea, dyspnea
• Palpitations
• Lightheadedness

Differential diagnosis for atrial fibrillation:

• MI
• MAT
• Atrial flutter
• Tachyardia

Etiology of atrial fibrillation:

• Hypertension
• Anemia
• Thyrotoxicosis
• COPD
• Idiopathic

And this may be followed by the memorization of a "typical" presentation of a patient with a particular disease and treatment:

A 57-year-old man with a history of hypertension presents to the emergency department with a 3-hour history of palpitations, shortness of breath, and lightheadedness. An EKG shows irregularly irregular rate and his blood pressure is 105/70.

What this creates is not expertise in medical practice, but storytellers great at just-so stories with no regards towards accuracy, noise reduction, and uncertainty. This process of learning medical information neglects the fact that probability of data (signs, symptoms, test results, etc) given a disease is not the same as the probability of a disease given the data. Presentation of information, as shown above, does not fit with a realistic framework that medical knowledge is uncertain and continuously changing, signs and symptoms vary, diagnostic tests are not perfect, and treatments are not 100% effective. This decontextualized framework of learning and assessing information is flawed. Furthermore, this creates an illusion of validity and overconfidence under the use of vague verbiage. Kahneman et al state the problem of not being aware of noise is that "people do not go through life imagining plausible alternatives to every judgment they make."⁵ This, on the other hand, is an important feature of the scientific method and the purpose of the differential diagnosis in medical practice if done well. But accuracy and noise reduction requires understanding how science works, including mastering the language of probability and uncertainty.

Acquiring more facts doesn't necessarily improve the accuracy of an intelligent system, as Steven Pinker reminds us it "must be equipped with a smaller list of core truths and a set of rules to deduce their implications."⁶ Information should not be viewed as neutral characteristic of the system used to pass exams and placed in books, they are part of the process of building our mental representations for accurate decision-making. Daniel Dennett claims that semantic "information is valuable—misinformation and disinformation are either pathologies or parasitic perversions of the default cases" adding that we "can’t be misinformed by distinctions we are not equipped to make."⁷ If education is not helping us to to develop a system to reduce noise, what is the purpose of education?

If information is a distinction that makes a difference, we are invited to ask: A difference to whom? Cui bono? Who benefits?

Daniel Dennett

Tweet to @jvrbntz

References:

1. Epstein, R., The empty brain: Your brain does not process information, retrieve knowledge or store memories. In short: your brain is not a computer., Aeon, May 18, 2016, Accessed November 20, 2017
2. Shannon, C. E. (1948), A Mathematical Theory of Communication. Bell System Technical Journal, 27: 379–423. doi:10.1002/j.1538-7305.1948.tb01338.x
3. "Information theory." Wikipedia: The Free Encyclopedia. Wikimedia Foundation, Inc., November 13, 2017, Accessed November 20, 2017
4. Gallistel, C. R., and Adam Philip. King. Memory and the Computational Brain: Why Cognitive Science will Transform Neuroscience. Hoboken, John Wiley & Sons, 2011
5. Kahneman, D., et al., Noise: How to Overcome the High, Hidden Cost of Inconsistent Decision Making., HBR, October 2016, Accessed November 20, 2017
6. Pinker, S., How the mind works., 1997
7. Dennett, D., From Bacteria to Bach and Back: The Evolutions of Minds., 2017

Image 2 source: https://xkcd.com/795/