How Aristotle Set in Motion the Bias in Machine Learning Models Today

by | Aug 18, 2021

1 min read

Machine learning in how Aristotle set in motion the bias in machine learning models with Twain Liu, founder of SENSEUS, a data classification system to improve NLU and universal coherency, by the MarsPod – How AI is Changing Society host by Mark Wesley, Marketing Manager, at MarsCrowd.

Our favorite quote:

Twain quote (1)

About this episode: Aristotle’s binary model and AI bias

As someone inquisitive about scientific truth, Twain kept “unknotting the string” until Aristotle. His binary system laid down a: syllogistic, dualistic, binary, and deterministic logic. “Something is either false (zero) or true (one)”.

Twain explains how Aristotle weaponized Pythagoras’ table of opposites. For example, rational or irrational, round or square, long or short. But when Aristotle “arrived”, he added male or female. Thus, he added something organic into a very rigid, polar-opposite structure.

Plus, Twain explains how Aristotle’s bias is ingrained from UX design and data analytics to systemic biases in datasets and algorithm designs.

What is Aristotle’s Logic?

Aristotle, the famed philosopher is the inventor of logic (logical reasoning). His logic system are under the lenses of “reasoning,” leading to apply his “syllogisms.” Syllogisms are deductions consisting of two types of argument: deduction and induction. Both, pointing to a conclusion.

Aristotle defines deduction as the “assumption of things” that make a premise, and the “results of the necessity” are the conclusion. Induction is an argument moved from the particular to the generalizations. In machine learning, machines use inductive and deductive logic to produce reason and learn. (See more at Stanford Enclyclopedia of Philosophy)

How does Aristotle’s logic created AI bias?

Machine learning algorithms are designed based on input data to provide a conclusion. Taking into account, Turing broke the separation between “machine, program, and data,” revolutionizing how these are not separated.

Moreover, the breakthrough of deep learning and neural networks mimicking human thought. Aristotle creates the foundational premise of the “law of thought” sustains Twain’s  argument about how humans replicate these biases and restrictive ways of logic into AI models.

This MarsPod episode also includes:

    • How to change machine learning methods not to represent Aristotle’s binary model.
    • Understand how quantum processing is (in Twain’s point of view) the way to avoid Aristotle’s bias.
    • How Twain aims to achieve equal representation in AI and not classify “1” as “male,” “good,” and “white”.

Listen to the full episode “How Aristotle Set in Motion the Bias in Machine Learning Models Today” on the MarsPod Podcast in all your favorite channels listed below. You’ll also find this episode on YouTube. Just search “MarsPod.”

amazonmusic-icon applepodcasts-icon googlepodcasts-icon spotify-icon

0 Comments

Submit a Comment

Your email address will not be published. Required fields are marked *

Like what you’re reading?
Get conversation tips straight to your inbox
Categories
Related Blog posts