Skip to content

llami-team/curiosity

Repository files navigation

⚗️ Curiosity (WIP)

This project is a project created to explore whether artificial embodiment of human consciousness is possible (like a Artificial Consciousness), not using conventional neural network system, but following the theory of expert system advocated by Marvin Minsky.

Abstract

Conventional artificial intelligence relies on the implementation of all intelligence on statistical models using artificial neural networks. But these conventional artificial intelligence are similar to simulating intelligence or consciousness by statistics, rather than exploring the principles of consciousness.

Recognizing and identifying things is the role of intelligence, and the category of consciousness that a person is experiencing lies in things like logic and emotion. In other words, this project was initiated under the assumption that consciousness and intelligence were in different categories.

Layer System

This project hierarchically classifies the minimum elements required to construct consciousness. (Like TCP/IP Layer)

Layer 0

It is assumed that layer 0 is constructed by a commonly called general-purpose computing machine, computer, or CPU.

  • Calculate

Layer 1

Layer 1 is the most basic step of consciousness to be developed in this project. Assume that the most basic and independent compositions of consciousness are constructed on the Layer 1.

  • Emotion
  • Logic
  • Language
  • Decision

Layer 2

Layer 2 acts like an ego of human consciousness. Each element of Layer 2 is constructed based on all the elements configured in Layer 1. For example, all elements of Layer 1 are used in the planning process.

  • Plan
  • Action
  • Caution
  • Expect
  • Judgment
  • Talk
  • Describe

Layer 3

Layer 3 serves as a tool for developing consciousness on its own and achieving its purpose. Each element of Layer 3 is constructed based on all the elements configured in Layer 2.

  • Learning
  • Recognition

Layer 1

Layer 1 consists of the most primitive and independent modules. However, they can refer to each other between Layer 1 elements as necessary.

Language

The language module aims to establish an architecture that can clearly and continuously define the association between words and words.

Language acts as a medium to make up information about any Essence or Existence.

Essence

The essence does not change or disappear, it is the most fundamental and unique information. In order for an Existence (such as a Subject or Object) to exist, the essence is essential.

Subcomponent of essence
State
  • Ex. The saying A is B is lie.
Type
  • Ex. A is B.
Relation
  • Ex. A is family with B.
Attributes of essence

The subcomponents of all essence may be limited by the following attributes:

Condition
  • Ex. A is only in the same age as B. (Relation + Conditions)

Existence

Existence is the concept of time added to Essence. Every time Essence changes, new Essence emerges as it changes, and all Essence is unique information that is individually distinct.

Subcomponent of Existence
  • Essences (Essences of the past that have changed over time)
Q. Why should all Essence be keep memory in Existence?

Essence is a unique being that cannot be changed. Just a new Essence is created and replaced within the Existence. Remembering this becomes information that remembers the state of the past. While it may be necessary not to remember the past artificially, additional development can be made only by remembering the past.

Q. Why are conditions present in the subcomponents of the essence?

The conditions that lead Existence to re-reference any other Essence shall be within Essence itself.

Q. Why is the language module unwilling to use existing classifications, such as conventional nouns or adjectives?

I think the reason why conventional Morpheme analysts couldn't be all-around is because they're obsessed with the relationship of superficial words. Before classifying the relationship of a word, I think that a step-by-step approach should be made to categorize the most fundamental and core minimum properties that the word points to, and then form a word system on it so that interpretation of the language by the program can take place.

Q. Is this language module aimed at Universal Grammar (UG)?

No, It is not possible to implement this module by itself. This module aims to find an essential configuration that logically separates the correlations of words.

Logical Learning plan

The goal is to find out how much basic logic is needed to enable basic logical learning through the translation and inductive processes through textual conversations with people.

  1. Learning about Essence
    • Define the subcomponents of Essence (status, type, relationship).
  2. Learning about Existence
    • Define the relationships of several Existence.

Logic

Series of processes to inferring an Essence from an Existence, or to define a new Existence from Essence, or to determine the true lie of a condition.

Deduction

The purpose is to verify the veracity of the argument, and if the premise is true, the result is true.

Summarize
  • A=B
  • B=C
  • C=A
Description
  • Principle: Human is a creature.
  • Case: Creatures eat food.
  • Result: Human should eat food.
Exception case
  • If the result is not correct
    1. There are insufficient conditions in the case,
    2. Either it's the wrong case.
    3. Either it's a wrong principle.
Example of adding conditions on exception
  • Case1 = Creatures eat food. (When it lacks nutrients)
  • Case2 = Creatures eat food. (Before nutrients are consumed excessively)

Induction

It is aimed at expanding knowledge beyond the argument of the premise. This allows some leap of logic in the process. An induction can create a probable hypothesis, but it does not have logical inevitability.

Summarize
  • A=D
  • B=D
  • C=D?
Description
  • Case 1= The cat on the street likes chur.
  • Case 2= The cat next door likes chur.
  • Inference = All cats will like chur.
Classification
  • When all cases have been validated

    • Complete induction
  • When some of the cases have not been validated

    • Imcomplete induction
  • Dialectic

Emotion​

Emotions are clearly in the process of defining what decisions are being made. Also emotions can also be classified as being made up of positive and negative models. and when the emotions are applied to an object, depending on the subcomponent of the subject, different emotions can act on one object simultaneously.

Emotion 1

+ Happiness

- Sadness

Emotion 2

+ Expectation

- Disappoint

Emotion 3

+ Love

- Anger

Emotion 4

+ Surprise

- Panic

Emotion 5

+ Comfort

- Anxiety

Emotion 6

+ Yearn

- Disgust

Emotion 7

+ Desire

- Shame

Decision

About

⚗️ Curiosity about consciousness

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published