File size: 4,855 Bytes
43c9a60
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
---
license: mit
task_categories:
  - time-series-forecasting
  - other
tags:
  - artificial-life
  - consciousness
  - evolutionary-computation
  - code-evolution
  - emergence
  - hard-problem
  - digital-organisms
  - ecosystem-dynamics
  - pandemic-simulation
language:
  - en
pretty_name: "Primordial: Artificial Life Evolution Dataset"
size_categories:
  - 1K<n<10K
---

# Primordial: Artificial Life Evolution Dataset

**Tick-by-tick evolution data from digital organisms with body (executable code) + mind (structured knowledge). Sexual reproduction, pandemics, mass extinction, predation — all emerged from simple rules.**

## Why This Dataset Exists

Most AI datasets capture static snapshots. This dataset captures **dynamic evolution** — digital organisms eating code, reproducing sexually, getting sick, dying, and being selected by nature over 2000 ticks.

No existing dataset provides:
1. **Ecosystem dynamics with full observability** — population, disease, energy, all tracked per tick
2. **Emergent phenomena from simple rules** — no behavior was hardcoded
3. **Sexual reproduction + pandemic cycles + mass extinction** in one simulation

## Dataset Contents

### `train.parquet` — Evolution Log (2000 records, Parquet format)

Also available as `sim3_evolution.jsonl` (raw JSONL).

One record per simulation tick. 60 initial entities, 2000 ticks.

**18 fields per record:**

| Field | Type | Description |
|-------|------|-------------|
| `tick` | int | Simulation step (1-2000) |
| `alive` | int | Living entities |
| `males` | int | Male count |
| `females` | int | Female count |
| `sick` | int | Currently infected |
| `max_gen` | int | Highest generation alive |
| `avg_energy` | float | Mean energy across population |
| `avg_body` | float | Mean body size (function count) |
| `avg_mind` | float | Mean mind size (knowledge nodes) |
| `max_body` | int | Largest body |
| `max_mind` | int | Largest mind |
| `avg_age` | float | Mean age in ticks |
| `born` | int | Births this tick |
| `died` | int | Deaths this tick |
| `eaten_pred` | int | Predation kills this tick |
| `mated` | int | Successful matings this tick |
| `infections` | int | New infections this tick |
| `famine` | bool | Famine period active |

### Key Statistics

| Metric | Value |
|--------|-------|
| Max generation | 22 |
| Total born | 4,785 |
| Total predation kills | 3,901 |
| Total infections | 6,447 |
| Peak population | 515 (tick 327) |
| Min population | 14 (tick 56) |
| Peak pandemic | 69% infected (tick 556) |
| Mass extinctions | 3 (ticks 600, 1200, 1800) |
| Carrying capacity | ~300-400 (emerged, not hardcoded) |

## Emergent Phenomena Captured

1. **Ecological oscillation** — population self-regulates around carrying capacity without any target
2. **Pandemic waves** — disease peaks then declines as immune memory spreads; new strains restart cycle
3. **Post-extinction recovery** — after 50% die-off, population rebounds within ~100 ticks; survivors are fitter
4. **Gender homeostasis** — M/F ratio stays ~50:50 despite stochastic assignment
5. **Generation acceleration then stabilization** — early gens appear fast, then stabilize to ~90-100 ticks/gen

## How To Use

### Load with HuggingFace
```python
from datasets import load_dataset
ds = load_dataset("jkdkr2439/Primordial-Evolution")
```

### Load directly
```python
import pandas as pd
df = pd.read_parquet("train.parquet")
df.plot(x="tick", y=["alive", "sick"], figsize=(12, 4))
```

### Prediction task
```python
# Predict next 100 ticks from previous 100
# Input: df[0:100], Output: df[100:200]
# Features: alive, sick, avg_energy, born, died, famine
```

## Connection to Consciousness Research

This dataset comes from the [Primordial](https://github.com/jkdkr2439/Primordial-Hard-Problem-of-Consciousness) project — a computational framework studying emergence of self-reflective behavior in digital organisms, with implications for the Hard Problem of Consciousness (Chalmers, 1995).

## Entity Architecture

```
Entity = Body (Python functions) + Mind (NMF knowledge graph)
  - Body: digest code, build new functions, decay unused ones
  - Mind: absorb knowledge, compress to DNA, decay unused nodes
  - Sex: M encodes gamete (cheap), F decodes + builds child (expensive)
  - Immune: antibody memory, virus corrupts body functions
  - Lifecycle: Vo (dormant) > Sinh (born) > Dan (growing) > Chuyen (transform)
```

## Physics Laws (all hardcoded, no entity can break)

- Syntax validity: `ast.parse()` — invalid code = dead
- Energy conservation: eating gives energy, existing costs energy
- Complexity cost: more code = more expensive to maintain
- Aging: older entities cost exponentially more
- Famine: every 250 ticks, food drops to 30%
- Extinction: every 600 ticks, bottom 50% killed

## Author

Tung Nguyen (Kevin T.N.)

## License

MIT