https://github.com/ajithksenthil/personalitymediatednarrativegen

Using Markov Chains and Computational Psychodynamics, a process-based approach to modeling event driven personality and cognition, in generative story worlds
https://github.com/ajithksenthil/personalitymediatednarrativegen

behavior cognition-and-perception generative-story-worlds gpt llm markov-chain narrative-generation personality

Last synced: 19 days ago
JSON representation

Using Markov Chains and Computational Psychodynamics, a process-based approach to modeling event driven personality and cognition, in generative story worlds

• Host: GitHub
• URL: https://github.com/ajithksenthil/personalitymediatednarrativegen
• Owner: ajithksenthil
• Created: 2023-09-13T00:19:21.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-11-01T18:35:46.000Z (about 1 year ago)
• Last Synced: 2023-11-01T19:29:53.135Z (about 1 year ago)
• Topics: behavior, cognition-and-perception, generative-story-worlds, gpt, llm, markov-chain, narrative-generation, personality
• Language: HTML
• Homepage:
• Size: 220 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Personality Mediated Generative Story Worlds

# Introduction

Drawing inspiration from computational linguistics and computational psychodynamics, this project dives into the realm of generative story worlds. At its core is the modeling of event-driven personality and cognition to create dynamic narratives.

# Computational Psychodynamics

This methodology bridges the gap between computational linguistics and behavior modeling. Just as computational linguistics combines linguistic schemas with statistical models, computational psychodynamics offers a structured schema for understanding behavior and cognition, using tools such as Markov chains and Transformer models. Key principles include:

Behavioral schema rooted in Jungian cognitive functions

The Free Energy Principle-Active Inference (FEP-ActInf) framework

Linearly separable binary classifications for behavioral states

Transition modeling between behavioral states over time

# Features

Dynamic Modeling: Using the principles of Computational Psychodynamics, this project captures the nuances of personality and behavior in generative story worlds.

AI Integration: The methodologies emphasize potential AI applications, especially in portraying behavior and personality through multi-level binary classifications.

Ecological View of Personality: Observations of behaviors form the main source of data, ensuring a more grounded and holistic understanding of personality dynamics.

# How It Works

Behavioral Schema: At the foundation is a behavioral schema based on Jungian cognitive functions, categorizing judgment/observer functions with perception/decider functions.

FEP-ActInf Integration: The schema is connected with the FEP-ActInf framework, emphasizing the brain's role in reducing differences between predicted and actual sensory inputs.

Transition Modeling: Behavioral transitions over time are modeled using Markov chains, Recurrent Neural Networks, and Transformer models.

Generative Narratives: The behavioral models are then utilized to create dynamic narratives, reflecting realistic personality-driven scenarios.

Applications

Brain-computer interfaces: Refining cognitive models for improved interaction and understanding.

AI personalities: Create more human-like AI entities with distinct personalities.

Generative narrative worlds: Craft intricate story worlds driven by dynamic characters.

Behavior prediction & analysis: Enhance the prediction and analysis of behavior over time.

# Related Manuscripts in Preparation

Computational Psychodynamics: Process-Based Framework for Modeling Cognition and Personality Using Active Inference: 

Delve deeper into the theory of Computational Psychodynamics and its applications in modeling event-driven personality and cognition in generative story worlds.