Kernel Based Cognitive Architecture for Autonomous Agents

Alexander Serov

DOI: https://doi.org/10.48550/arXiv.2207.00822

2022-07-02

Abstract:One of the main problems of modern cognitive architectures is an excessively schematic approach to modeling the processes of cognitive activity. It does not allow the creation of a universal architecture that would be capable of reproducing mental functions without using a predetermined set of perceptual patterns. This paper considers an evolutionary approach to creating a cognitive functionality. The basis of our approach is the use of the functional kernel which consistently generates the intellectual functions of an autonomous agent. We consider a cognitive architecture which ensures the evolution of the agent on the basis of Symbol Emergence Problem solution. Evolution of cognitive abilities of the agent is described on the basis of the theory of constructivism.

Artificial Intelligence

What problem does this paper attempt to address?

The paper attempts to address the issue of overly simplified modeling methods in modern cognitive architectures, which cannot create a universal architecture capable of reproducing psychological functions without using preset perception modes. Specifically, the authors focus on how to generate cognitive functions of autonomous agents through a functional kernel and ensure the cognitive evolution of the agent throughout its entire lifecycle. This process is based on solving the Symbol Emergence Problem (SEP) and describes the evolution of the agent's cognitive abilities through constructivism theory.

### Main Issues of the Paper:
1. **Limitations of Modern Cognitive Architectures**:
- Modern cognitive architectures use overly simplified modeling methods, making it impossible to create a universal architecture capable of reproducing psychological functions.
- There is a lack of a bottom-up autonomous learning method, resulting in insufficient methodologies to describe such tasks.

2. **Symbol Emergence Problem (SEP)**:
- Modern biologically inspired cognitive architecture (BICA) projects are overly "mechanized" in representing cognitive processes, mainly due to the failure to solve the symbol emergence problem.
- The core of the symbol emergence problem lies in how to create basic symbols of a system that do not rely on predefined semantic components.

3. **Constructing a Universal Functional Kernel**:
- The authors propose a universal cognitive architecture based on a functional kernel that ensures the cognitive evolution of the agent throughout its entire lifecycle.
- The goal of the functional kernel is to maximize the agent's survival probability and initiate cognitive processes.

### Key Points of the Solution:
- **Artificial Subjective Reality**:
- Using the concept of semiotics to describe the agent's subjective world (Umwelt), including the perceptual world (Merkwelt) and the operational world (Werkwelt).
- Simulating the development process of cognitive functions through a constructivist AI approach.

- **Design of the Functional Kernel**:
- The functional kernel is divided into two levels: the Umwelt interface layer directly related to the agent entity and the psychological function layer related to advanced psychological functions.
- The kernel contains a series of algorithms that are activated immediately after the agent's birth, ensuring the agent's survival and the initiation of cognitive processes.

- **Components of the Cognitive Architecture**:
- Perceptive component: Provides data about the Merkwelt.
- Motor component: Responsible for low-level control and information transmission.
- Intelligent component: Creates, modifies, stores, and manages psychological models (symbols) that describe reality.
- Emotional component: Involved in learning and pattern recognition, helping to create and classify high-level psychological models.
- Volitional component: Coordinates the work of other core components and allocates energy resources.

### Conclusion:
The paper aims to solve the simplified modeling problem and the symbol emergence problem in modern cognitive architectures by constructing a universal cognitive architecture based on a functional kernel. This architecture ensures the cognitive evolution of the agent throughout its entire lifecycle, achieving higher levels of autonomy and intelligence.