5000 Years of Modern Computing

Deductive reasoning

systematically argue from your evidence.

Plan a computation with the right logic, like the rules of a game.

Identify the logical notions in a context by focusing one does not change, like the flagpole not the flag.

Choose the right type for your data so that all the tools that use it are the best for the job.

Identify membership of data in subsets and create equivalence between members through partitions like membership categories.

systematically work beyond your evidence

Structure data and arguments around cases and patterns.

Use relations to convert data into new propositions.

Make efficient logical deductions using defaults, like assuming a bird can fly, and recover to errors using adaptive and dynamic logic.

Make informed guesses and measure the likelihood of a proposition.

Argue to the best hypothesis of your evidence

Identify the correct meaning of proximity so that your measurements get closer to the destination and do not wonder off course.

Create measurements so that the whole is the sum of its parts allowing it become a data feature and used in machine learning.

Project data to single numbers called “features” by treating features as integrals.

Machine learn by taking a shot at a feature, measure the error, and back propagate the correction through derivatives.