Spaces:
Runtime error
Runtime error
File size: 10,248 Bytes
12af533 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 | #!/usr/bin/env python3
"""
Layer 0: Core Mathematical Framework
Based on math-skills.txt and skill-ecosystem-integration-guide.md
This is the FOUNDATION - all other layers build on this.
"""
import numpy as np
from typing import List, Dict, Tuple, Optional
from dataclasses import dataclass, field
import json
# ============================================================================
# SKILL REPRESENTATION (From math-skills.txt)
# ============================================================================
@dataclass
class Skill:
"""
8-dimensional skill representation in ℝ⁸
s = (G, C, S, A, H, V, P, T) ∈ ℝ⁸
"""
name: str
# 8 dimensions with weights
G: float = 0.0 # Grounding (18%)
C: float = 0.0 # Certainty (20%)
S: float = 0.0 # Structure (18%)
A: float = 0.0 # Applicability (16%)
H: float = 0.0 # Coherence (12%)
V: float = 0.0 # Generativity (8%)
P: float = 0.0 # Presentation (5%)
T: float = 0.0 # Temporal (3%)
# Metadata
priority: float = 0.5
cost: float = 1.0
embedding: Optional[np.ndarray] = None
def __post_init__(self):
if self.embedding is None:
# Create embedding from 8 dimensions
self.embedding = self.to_vector()
def to_vector(self) -> np.ndarray:
"""Convert to vector representation"""
return np.array([self.G, self.C, self.S, self.A, self.H, self.V, self.P, self.T])
def q_score(self) -> float:
"""
Calculate Q-score:
Q = 0.18×G + 0.20×C + 0.18×S + 0.16×A + 0.12×H + 0.08×V + 0.05×P + 0.03×T
"""
return (
0.18 * self.G +
0.20 * self.C +
0.18 * self.S +
0.16 * self.A +
0.12 * self.H +
0.08 * self.V +
0.05 * self.P +
0.03 * self.T
)
# ============================================================================
# SIMILARITY FUNCTIONS (From math-skills.txt)
# ============================================================================
def cosine_similarity(v1: np.ndarray, v2: np.ndarray) -> float:
"""
Cosine similarity:
sim(vᵢ, vⱼ) = (vᵢ · vⱼ) / (||vᵢ|| · ||vⱼ||)
"""
dot_product = np.dot(v1, v2)
norm_product = np.linalg.norm(v1) * np.linalg.norm(v2)
if norm_product == 0:
return 0.0
return dot_product / norm_product
def build_adjacency_matrix(skills: List[Skill]) -> np.ndarray:
"""
Adjacency matrix A where:
A_ij = sim(vᵢ, vⱼ)
"""
n = len(skills)
A = np.zeros((n, n))
for i in range(n):
for j in range(n):
A[i, j] = cosine_similarity(skills[i].embedding, skills[j].embedding)
return A
# ============================================================================
# INTERACTION TENSOR (From skill-ecosystem-integration-guide.md)
# ============================================================================
def compute_interaction(s1: Skill, s2: Skill, alpha: float = 0.7, beta: float = 0.3) -> float:
"""
Interaction strength:
I(i,j,k) = α · sim(vᵢ,vⱼ) + β · ∇²ₖE(sᵢ, sⱼ)
Classification:
- I > 0.7: SYNERGISTIC
- |I| < 0.3: INDEPENDENT
- I < -0.5: ANTAGONISTIC
"""
sim = cosine_similarity(s1.embedding, s2.embedding)
# Simplified curvature term (for now just use Q-score difference)
q_diff = abs(s1.q_score() - s2.q_score())
curvature = -q_diff # Negative because large difference = antagonistic
interaction = alpha * sim + beta * curvature
return interaction
def classify_interaction(interaction: float) -> str:
"""Classify interaction type"""
if interaction > 0.7:
return "SYNERGISTIC"
elif abs(interaction) < 0.3:
return "INDEPENDENT"
elif interaction < -0.5:
return "ANTAGONISTIC"
else:
return "NEUTRAL"
# ============================================================================
# SKILL SYNTHESIS (From math-skills.txt)
# ============================================================================
def synthesize_skills(parents: List[Skill], gamma: float = 0.20) -> Skill:
"""
Emergent skill synthesis:
s_emergent = Σᵢ αᵢsᵢ + γ · (s₁ ⊗ s₂ ⊗ ... ⊗ sₖ)
For now: weighted average + tensor product term
"""
# Calculate weights based on Q-scores
q_scores = np.array([s.q_score() for s in parents])
weights = q_scores / q_scores.sum()
# Weighted average of dimensions
vectors = np.array([s.to_vector() for s in parents])
avg_vector = np.average(vectors, axis=0, weights=weights)
# Tensor product term (simplified: element-wise product of normalized vectors)
normalized_vectors = vectors / np.linalg.norm(vectors, axis=1, keepdims=True)
tensor_term = np.prod(normalized_vectors, axis=0) * gamma
# Combine
emergent_vector = avg_vector + tensor_term
# Clip to [0, 1]
emergent_vector = np.clip(emergent_vector, 0, 1)
# Create emergent skill
emergent = Skill(
name="+".join([s.name for s in parents]),
G=emergent_vector[0],
C=emergent_vector[1],
S=emergent_vector[2],
A=emergent_vector[3],
H=emergent_vector[4],
V=emergent_vector[5],
P=emergent_vector[6],
T=emergent_vector[7],
priority=max(s.priority for s in parents),
cost=sum(s.cost for s in parents) / len(parents)
)
return emergent
# ============================================================================
# UTILITY CALCULATION (From math-skills.txt)
# ============================================================================
def compute_utility(skill: Skill, task_embedding: np.ndarray) -> float:
"""
Utility of skill for task:
U_i(t) = q_i · g(sim(v_i, t))
where g(s) = max(0, s) for simplicity
"""
sim = cosine_similarity(skill.embedding, task_embedding)
g = max(0, sim) # Activation function
utility = skill.q_score() * g
return utility
# ============================================================================
# SKILL SELECTION (From math-skills.txt)
# ============================================================================
def select_skills(
skills: List[Skill],
task_embedding: np.ndarray,
budget: float = 10.0,
gamma: float = 0.1
) -> List[int]:
"""
Greedy skill selection:
max_x x^T U + (γ/2) x^T A x subject to c^T x ≤ B
Greedy approximation: rank by marginal gain / cost
"""
n = len(skills)
selected = []
selected_indices = []
total_cost = 0.0
# Build adjacency matrix
A = build_adjacency_matrix(skills)
# Compute utilities
utilities = np.array([compute_utility(s, task_embedding) for s in skills])
# Greedy selection
remaining = list(range(n))
while remaining and total_cost < budget:
best_gain = -np.inf
best_idx = None
for i in remaining:
# Marginal utility
u_i = utilities[i]
# Synergy with selected skills
synergy = 0
for j in selected_indices:
synergy += A[i, j]
synergy *= gamma
# Total marginal gain
marginal_gain = u_i + synergy
# Per-cost gain
gain_per_cost = marginal_gain / skills[i].cost
if gain_per_cost > best_gain:
best_gain = gain_per_cost
best_idx = i
if best_idx is None:
break
# Add if within budget
if total_cost + skills[best_idx].cost <= budget:
selected_indices.append(best_idx)
selected.append(skills[best_idx])
total_cost += skills[best_idx].cost
remaining.remove(best_idx)
else:
break
return selected_indices
# ============================================================================
# TESTING
# ============================================================================
if __name__ == "__main__":
print("=" * 60)
print("LAYER 0: MATHEMATICAL FOUNDATION TESTS")
print("=" * 60)
# Create test skills from research documents
meta_learning = Skill("Meta-Learning", G=0.95, C=0.92, S=0.95, A=0.98, H=0.95, V=0.92, P=0.90, T=0.90)
transfer = Skill("Transfer-Learning", G=0.95, C=0.92, S=0.95, A=0.98, H=0.95, V=0.92, P=0.90, T=0.90)
problem_solving = Skill("Universal-PS", G=0.95, C=0.92, S=0.95, A=0.98, H=0.95, V=0.92, P=0.90, T=0.90)
print("\n[Test 1] Q-Score Calculation")
print(f" Meta-Learning Q: {meta_learning.q_score():.3f}")
print(f" Transfer Q: {transfer.q_score():.3f}")
print(f" Problem-Solving Q: {problem_solving.q_score():.3f}")
print("\n[Test 2] Similarity Matrix")
skills = [meta_learning, transfer, problem_solving]
A = build_adjacency_matrix(skills)
print(f" Adjacency matrix:\n{A}")
print("\n[Test 3] Interaction Classification")
interaction = compute_interaction(meta_learning, transfer)
classification = classify_interaction(interaction)
print(f" Meta-Learning ↔ Transfer: {interaction:.3f} ({classification})")
print("\n[Test 4] Skill Synthesis")
emergent = synthesize_skills([meta_learning, transfer])
print(f" Parents: {[s.name for s in [meta_learning, transfer]]}")
print(f" Parent Q avg: {np.mean([s.q_score() for s in [meta_learning, transfer]]):.3f}")
print(f" Emergent: {emergent.name}")
print(f" Emergent Q: {emergent.q_score():.3f}")
print(f" δ (gain): {emergent.q_score() - np.mean([s.q_score() for s in [meta_learning, transfer]]):.4f}")
print("\n[Test 5] Skill Selection")
task = np.array([0.9, 0.9, 0.8, 0.95, 0.85, 0.8, 0.75, 0.7]) # Task embedding
selected_indices = select_skills(skills, task, budget=5.0)
print(f" Selected skills: {[skills[i].name for i in selected_indices]}")
print("\n" + "=" * 60)
print("ALL LAYER 0 TESTS COMPLETE")
print("=" * 60)
|