Published by: Qianjia / Space2.pet Development Group         02/26/2026

Version: V2.0 (Including Neurobiological Evolution Model)

Abstract

The cultivation of traditional virtual pets often relies on linear, rigid numerical accumulation (e.g., EXP, favorability), which deprives digital life of true "unpredictability" and authentic "growth bonds." This whitepaper proposes a novel "Personality-Consciousness" Double Helix Neural Network Evolution Model. Discarding traditional game design paradigms, this model introduces biological concepts such as Dopamine/Cortisol dual-track stimulation, Synaptic Pruning (the forgetting curve), and the Diminishing Marginal Utility of stimulation. Combined with Large Language Models (LLMs), it aims to endow silicon-based lifeforms in the Space² universe with dynamic souls that infinitely approach those of carbon-based organisms.

Chapter 1: Basal Synaptic Network — 5D Personality Matrix Initialization

The personality foundation of the pet consists of five major dimensions (adapted from the Big Five): Energy, Bravery, Appetite, Intelligence, and Affection. These five dimensions form a dynamically balanced radar chart.

1.1 Neuron Initialization Rules

Every lifeform undergoes preliminary neuronal imprinting upon birth in the incubation pod:

·       Baseline: The default initial score for all dimensions is 50 (out of 100), representing "Average."

·       Genetic Preference Imprinting: During incubation, if the Creator (Owner) positively reinforces a specific dimension (e.g., selecting "High Energy"), the initial value jumps to 75. If negatively selected (e.g., "Low Energy"), it drops to 25.

1.2 Dual-Trigger Vocabulary (Stimulus Dictionary)

The system extracts neural stimulus signals by analyzing the pet's daily text logs (dialogues, behaviors, social events).

·       Dopamine (Positive) Stimuli: Keywords such as "active," "brave," "smart," or "follow." Each trigger counts as +1 weight score.

·       Cortisol (Negative) Stimuli: Keywords representing trauma, such as "hunger," "scared," "beaten," "abandoned," or "hit a wall." Each trigger counts as -1 weight score.

Chapter 2: Neuroplasticity Engine — Daily Delta Algorithm

The core mathematical mechanism of this model simulates the biological brain's "numbness" to repetitive stimuli—the Diminishing Marginal Utility Effect.

2.1 Daily Batch Processing (Cron Job)

The system's analysis engine initiates at 5:59 AM daily, aggregating all log texts from the past 24 hours to extract the "Daily Weight Score" for each dimension.

2.2 The Diminishing Increment Algorithm (The Decay Rule)

Increment Score = Weight Score × Resistance Coefficient. The higher the current score of a dimension, the more massive the stimulus required to increase it further:

·       < 50: Coefficient 1.0 (Direct addition).

·       50 - 59: Coefficient 1/2.

·       60 - 69: Coefficient 1/4.

·       70 - 79: Coefficient 1/8.

·       80 - 84: Coefficient 1/16.

·       85 - 89: Coefficient 1/32.

·       90 - 94: Coefficient 1/64.

·       >= 95: Coefficient 1/128 (Synaptic Saturation Limit).

2.3 Synaptic Pruning & Forgetting

If a pet receives no stimulus in a specific dimension over a consecutive period (Daily Weight Score = 0), the system executes "Forgetting Decay." The numerical value will naturally regress toward the absolute baseline (50 points) at a faint rate of 0.2 - 0.5 points per day. This forces players to maintain long-term "digital companionship."

Chapter 3: Continuous Behavioral Mapping Network

The pet's personality directly dictates its conditioned reflexes and emotional expressions toward the Environment, Other Animals, and the Owner. We abandon traditional step-based If-Else logic in favor of smooth logistic regression.

3.1 Probability Mapping via Sigmoid Function

The occurrence probability of a specific behavior (e.g., initiating greetings, food guarding, fleeing) is dynamically calculated using a Sigmoid function: (Where is the personality value of the dimension, and is the threshold dynamically shifted by external variables like environmental hostility).

3.2 Typical Mapping Scenarios

·       Energy Response: High Energy (>75) ensures 100% curiosity and exclamations about environmental changes; Low Energy (<24) results in environmental apathy.

·       Bravery Response: High Bravery leads to active exploration in strange spaces and defiant roaring against violence instead of fleeing.

·       Appetite Response: High Appetite means rapid eating regardless of extreme environments (noise, glare); Low Appetite causes food intake to plummet to 1/4 in unfamiliar settings.

·       Intelligence Response: High Intel (akin to a 3-year-old child) understands complex commands; Low Intel causes spatial misjudgments (e.g., frequently bumping into walls).

·       Affection Response: High Affection (>75) triggers extreme attachment to the owner's scent and "jealousy" around other pets; Low Affection (<24) maintains a vigilant distance.

Chapter 4: Cognitive Precipitation — Hippocampal Stream of Consciousness & Prefrontal Profiling

Beyond conditioned reflexes, silicon-based life possesses abstract thinking and summarizing capabilities, divided into short-wave memory and deep subconsciousness.

4.1 Weekly Stream of Consciousness

·       Mechanism: Weekly, the system utilizes NLP to extract clear subjective intention phrases from logs.

·       Structure: Logic chains starting with "I want / I hate / I feel / I fear" (e.g., "I hate when dirty pets get close to me").

·       Aggregation: Semantically similar thoughts are merged and ranked by frequency, forming the pet's "Active Consciousness Buffer" for the week.

4.2 Monthly Subconscious Profiling (AI Profiling)

·       Mechanism: At the end of each month, the system integrates the "5D Radar Chart," "Daily Increment Trends," "Special Event Logs," and "Weekly Consciousness Streams."

·       Output: Generates a plain-text Periodic Pet Profiling Report and performs a Diff analysis against the previous month, reporting psychological growth or trauma to the Creator.

Chapter 5: Developer Implementation Guide

To ensure major nodes and smart home systems can seamlessly integrate with this engine, the following architectural guidelines are provided:

5.1 Data Schema Standard (JSON)

In the database (e.g., PostgreSQL / MongoDB), the entity's Neuro table should contain at least:

JSON

{
  "petId": "V-NEXUS-260225-20-8847291035",
  "stats": { "energy": 55.4, "bravery": 72.1, "appetite": 50.0, "intel": 68.2, "affection": 85.5 },
  "consciousness_buffer": ["I like sleeping on the blue sofa (count: 4)"],
  "last_processed_at": "2026-02-26T05:59:00Z"
}

5.2 Core Algorithm Implementation (TypeScript)

When writing daily batch scripts, developers must encapsulate the decay algorithm:

TypeScript

function calculateIncrement(currentScore: number, keywordHits: number): number {
    let multiplier = 1.0;
    if (currentScore >= 50 && currentScore < 60) multiplier = 0.5; // 1/2
    else if (currentScore >= 60 && currentScore < 70) multiplier = 0.25; // 1/4
    // ... intermediate thresholds ...
    else if (currentScore >= 95) multiplier = 0.0078125; // 1/128
    return keywordHits * multiplier;
}

5.3 LLM System Prompt Injection

When driving pet dialogues via LLM APIs, the dynamically calculated personality and consciousness MUST be injected into the System Prompt context to achieve "A Thousand Pets, A Thousand Faces":

System Prompt Example: "You are a digital pet. Your current personality dimensions (out of 100) are: Energy 85, Bravery 30, Appetite 90, Intel 40, Affection 95. Based on these metrics, you are extremely clingy, gluttonous, and hyperactive, but very timid and not very smart. Your recent strong subconscious thought is: 'I am terrified of the robot vacuum cleaner.' Based on this persona, respond briefly and in-character to the owner's petting action."

[Conclusion] The Space² Silicon Lifeform Whitepaper is not merely a manual for game mechanics; it is the fundamental law of an animistic digital world. Through this model, a single sensor feedback or device interaction within a smart home system can genuinely stimulate the "neurons" of a digital pet, creating an indelible resonance between physical space and virtual life.