Knowing

Tower of Babylon

#religious-fiction

#divine-retribution-motif

Sample Use Cases

A random process is ergodic if all of its statistics can be determined from a sample function of the process. That is, the ensemble averages equal the corresponding time averages with probability one.

Role of Ergodicity in Human Inference

A newspaper has previously printed some inaccurate information, therefore, the newspaper is going to publish inaccurate information in the future. Fair; ensemble of published articles is more or less ergodic.

More crimes are committed by black persons than by white persons, therefore each individual black person is not to be trusted? The ensemble of black people is not at all ergodic!

Mental Attitude While Learning

Distinguish Mere Facts From Conclusions or Opinions

Discriminate between mere statements of facts, necessary conclusions which follow therefrom, and mere opinions which they seem to render reasonable. There’s no need to perform an experiment to verify that the atomic weight of oxygen is 16. That the sum of the angles of a plane triangle equals two right angles is not a mere fact, but an inevitable truth.

Why Even Read Papers?

While tutorials and docs help you write code right now, the academic papers can help you understand where programming came from and where it’s going. You also understand the paths that foundational academic research did not take. There’s a lot of things that are old that are new again, over and over and over. The idea of Stack Overflow is that someone else has had your problem before; the idea of academic papers is that someone else has thought about this problem before.

The Relativity of Wrong. Isaac Asimov. chem.tufts.edu . Sep 1, 1989.

John, when people thought the earth was flat, they were wrong. When people thought the earth was spherical, they were wrong. But if you think that thinking the earth is spherical is just as wrong as thinking the earth is flat, then your view is wronger than both of them put together.

Brainless Slime That Can Learn by Fusing. Ed Yong. www.theatlantic.com . old.reddit.com . Dec 21, 2016.

Building Transit Networks

Can a cell learn? When a part of the plasmodium touches something attractive, e.g. food, it pulses more quickly and widens. If a part meets something repulsive, like light, it pulses more slowly and shrinks.

Norms for Scientific Activity and the Sources of Error

Internal norms derive from a cognitive pursuit of science. They are:

• Focus on the central rather than the peripheral problems of the science in in question.
• Craftsmanship standards in the design of experiments and hypothesis testing.
• Distinguished theorizing is elegant and simple.
• Searching for a theory that is closer to the truth than the currently accepted theory.

External norms govern motives for participating in and using the results of scientific activity, e.g.

Immeasurably Important. Morgan Housel. www.collaborativefund.com . Jul 5, 2018.

Filtering out information is an art, not a science, necessitated by the information overload that we live in.

• Watch out for the tendency to only preserve information that meshes with how we think the world should be.
• If you think the world is all art, you miss how much stuff is too complicated to think about intuitively.

If you think the world is all data, you’ll miss how much is too complicated to summarize in a statistic.

The Four Color Theorem does not claim that 4 colors suffice to color a planar map. Instead, 4 colors are sufficient to color any planar graph so that no two vertices connected by an edge are colored with the same color. For any \(n\), there is a map that requires at least \(n\) colors.

Similarity-based learning is intuitive and gives people confidence in the model.

There is an inductive bias that instances that have similar descriptive features belong to the same class.

Similarity learning has a stationary assumption, i.e. the joint PDF of the data doesn’t change (new classifications do not come up). This assumption is shared by supervised ML.

Furthermore, an NN model can only give answers that are present in the training set. Ergo, is your training set representative?

Similarity-based learning is intuitive and gives people confidence in the model.

There is an inductive bias that instances that have similar descriptive features belong to the same class.

Similarity learning has a stationary assumption, i.e. the joint PDF of the data doesn’t change (new classifications do not come up). This assumption is shared by supervised ML.

Furthermore, an NN model can only give answers that are present in the training set. Ergo, is your training set representative?