kant-what-would-you-do.md

Kant, what would you do?

Yet again, the machine learning (ML) & computer vision (CV) community, ever thirsty for "open science," exchanged a bunch of ideas on Twitter during the past week. The topic: CVPR reminding its community of the policy approved by them in June. Namely, the social-media ban while articles are under review.

Unfortunately, I was busy and could not exchange 240 chars thoughts with people who did not disclose participation in CVPR'21 motion, but are thirsty to share opinions (need attention? 🤔). As it's a weekend for writing, here we go.

I'm fortunate to have a good memory for stuff that I don't use. An example of that, it's the concept of "Categorical imperative" from Kant's ethics. I heard about it once back in 2007/2008, and rarely use it after that. I guess that I had an eloquent teacher that persuaded me to listen to him, Lecturer in philosophy Carlos Pajaro.

Instead of getting you bored with stuff that you can read from a knowledgeable source, I will explain it with a premise and two situations and questions.

• Kant's philosophy strikes for rationalism, and his thoughts about ethics and morals are not an exception.

Premise done! Now, the two situations:

1. Assuming that person-A breaks into your house scared, begging you for help for hiding. Would you concede so?

   1. Let's assume that you acknowledge that. Minutes later, another individual, person-B, knocks on your door and asks you: "have you seen a subject, with a description similar to person-A?"

   2. You are a rational person extremely attentive and perceive that person-B has a gun loaded. You also managed to sense that the person is infuriated. Thus, what would you do?

   You tell him the place in your house where person-A is hiding?

   
   Oops. If I'm person-A, I hope to not break in your place 🤣
   You ask person-B, why are you looking to that person?

   
   Followed by the rational question: "I notice that your gun is loaded. Would you hit an unarmed person?".

   Here, the rational ethical code should trigger to protect the life of person-A.

   😲 isn't it similar to the opening scenes for Unglorious Bastard? I thought Quentin was just thirsty for violence.

   

   Part-1 (Youtube) in case you wanna remember it.

2. Assume that we live in a society where it's culturally accepted to pee on a given wall.

   Fortunately (or unfortunately), you belong to an elite class of society. Those who know that peeing on the wall will damage it.

   ⁉️ Would you pee on the wall if you are sure that there are NO elite seeing you?

   Here, the rational ethical code should trigger saying: "What would it happen if all my elite fellows concede and pee on the wall?"

Those are the situations that I've used to remember the categorical imperative. I forgot the rest of the fat and only retained the meat. BTW, pardon me if I'm confusing Kant's ethics with someone else or the concept of the categorical imperative is more profound. As I said, I rarely have to refer to it.

Cute, but how is that related with people demanding the need of posting about their work in social media?

Are we peeing on the double-blind wall established by the community to hedge ourselves against our personal biases?

Instead of discussing the pros & cons of posting on social media as a "puritan" act protecting double-blind, I will appeal for the 1st situation:

• Have you asked the under-represented community fellows why they don't submit to CVPR/ICCV/ECCV?

  Maybe, you could think: "well there are more stupid scientists let's say in Colombian than in EEUU & Europe." Judging by the one writing this post, you have given a valid hypothesis 🤭.

• Did you know that in some countries conferences' proceedings like CVPR, yep the one chasing the throne of Nature 👑😋, count zero as a scientific contribution?

• How much do you think that arxiv counts?

Are you ready to hand over your under-represented fellows (person-A) to the punisher (person-B)?

CVPR/ICCV/ECCV is your house. Let's build a wall, the under-represented fellows will pay it!