AEAO Tetrad Configuration

Configure the four pillars of Sematryx's AEAO Engine: Agentic, Expository, Autodidactic, and Domain Extension. Control compute cost vs capability trade-offs with granular configuration options.

The AEAO Tetrad

Sematryx's AEAO Engine is built on four central pillars that work together to provide enterprise-grade optimization:

🤖

Agentic Intelligence

Multi-agent coordination for strategy selection. Research agents, validation engineers, and performance analysts collaborate to provide consensus-based optimization strategies.

  • • Multi-agent collaboration
  • • Consensus-based strategy selection
  • • Real-time performance analysis
  • • Autonomous decision-making
đź“–

Expository Intelligence

Comprehensive explanation of all optimization decisions with configurable explanation levels (0-5) for compute cost control.

  • • Configurable explanation levels (0-5)
  • • Natural language summaries
  • • Technical decision logs
  • • Interactive visualizations
  • • 22-26% performance boost with async processing
🧠

Autodidactic Intelligence

Self-improvement and continuous learning from optimization experience. Problem signature detection, strategy variation, and cross-problem learning.

  • • Problem signature detection
  • • Strategy variation
  • • Cross-problem learning
  • • Performance memory
  • • Meta-learning capabilities
🏗️

Domain Extension

Business domain libraries for rapid enterprise adoption. Engine-domain separation enables core optimization algorithms across 13+ business domains.

  • • 13+ business domains
  • • Automatic code generation
  • • FastAPI deployment
  • • Pattern recommendations
  • • Production-ready templates

Preset Configurations

Quick-start with predefined configurations optimized for different use cases:

Using Preset Configurations
from aeao import aeao

# Use preset configuration
result = aeao(
    objective_function=sphere,
    bounds=[[-5, 5], [-5, 5]],
    preset="production"  # development, production, research, enterprise, minimal
)

Available Presets

development: Fast iteration, basic explanations, minimal overhead. Best for rapid prototyping and testing.
production: Balanced performance, standard explanations, enterprise monitoring. Recommended for most production deployments.
research: Maximum capabilities, comprehensive explanations, full learning enabled. Ideal for research and experimentation.
enterprise: Full features, advanced monitoring, compliance features, audit trails. For regulated industries.
minimal: Core optimization only, no AI systems, fastest execution. For simple problems.

Simple Configuration

Enable specific tetrad pillars with simple boolean flags and basic parameters:

Simple Tetrad Configuration
from aeao import aeao

# Enable specific tetrad pillars
result = aeao(
    objective_function=sphere,
    bounds=[[-5, 5], [-5, 5]],
    use_agentic_intelligence=True,
    use_autodidactic_intelligence=True,
    explanation_level=3
)

Simple Configuration Options

  • use_agentic_intelligence (bool): Enable multi-agent coordination
  • use_expository_intelligence (bool): Enable explainability
  • use_autodidactic_intelligence (bool): Enable learning system
  • use_domain_extension (bool): Enable domain libraries (default: True)
  • explanation_level (int, 0-5): Detail level for explanations (0=off, 5=comprehensive)

Advanced Configuration

Fine-tune every aspect of the Tetrad with complete configuration objects:

Advanced Tetrad Configuration
from sematryx import sematryx, AEAOTetradCompleteConfig

# Complete custom configuration
config = AEAOTetradCompleteConfig.enterprise()
config.expository.explanation_level = 4
config.agentic.max_agents_per_problem = 5
config.autodidactic.learning_enabled = True
config.domain_extension.use_domain_libraries = True

result = aeao(
    objective_function=sphere,
    bounds=[[-5, 5], [-5, 5]],
    config=config
)

Advanced Configuration Options

Agentic Configuration:
  • • max_agents_per_problem: Maximum number of agents (default: 3)
  • • consensus_threshold: Agreement threshold for strategy selection
  • • agent_timeout: Timeout for agent responses
Expository Configuration:
  • • explanation_level: 0-5 detail level
  • • async_explanations: Background processing (default: True)
  • • include_visualizations: Generate visual diagnostics
  • • natural_language: Enable NLP summaries
Autodidactic Configuration:
  • • learning_enabled: Enable learning system
  • • cross_problem_learning: Learn across different problems
  • • memory_retention: How long to retain learned patterns
  • • meta_learning: Enable meta-learning capabilities
Domain Extension Configuration:
  • • use_domain_libraries: Enable domain-specific libraries
  • • auto_code_generation: Generate domain code automatically
  • • domain_patterns: Enable pattern recommendations

REST API Configuration

Configure the Tetrad via REST API requests:

REST API - Tetrad Configuration
curl -X POST https://api.sematryx.com/v1/optimize \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "objective_function_id": "func_1234567890",
    "bounds": [[-10, 10], [-10, 10]],
    "max_evaluations": 2000,
    "preset": "production",
    "tetrad_config": {
      "use_agentic_intelligence": true,
      "use_expository_intelligence": true,
      "use_autodidactic_intelligence": true,
      "use_domain_extension": true,
      "explanation_level": 3,
      "agentic": {
        "max_agents_per_problem": 5
      },
      "autodidactic": {
        "learning_enabled": true,
        "cross_problem_learning": true
      }
    }
  }'

JavaScript SDK Configuration

Configure the Tetrad using the JavaScript SDK:

JavaScript SDK - Tetrad Configuration
import { Sematryx } from '@sematryx/javascript-sdk'

const sematryx = new Sematryx('your-api-key')

// Option 1: Use preset
const result = await aeao.optimize({
  objective_function: sphere,
  bounds: [[-5, 5], [-5, 5]],
  preset: 'production'
})

// Option 2: Custom tetrad config
const result = await aeao.optimize({
  objective_function: sphere,
  bounds: [[-5, 5], [-5, 5]],
  tetrad: {
    use_agentic_intelligence: true,
    use_expository_intelligence: true,
    use_autodidactic_intelligence: true,
    use_domain_extension: true
  },
  expository: {
    explanation_level: 4
  },
  agentic: {
    max_agents_per_problem: 5
  }
})

Compute Cost vs Capability Trade-offs

Each Tetrad pillar can be independently enabled or disabled to balance performance and capabilities:

Minimal Cost (minimal preset)

Fastest execution, no AI overhead:

  • • All Tetrad pillars disabled
  • • Core optimization only
  • • ~10-20% faster than full Tetrad

Balanced (production preset)

Good performance with essential capabilities:

  • • Agentic + Expository enabled
  • • Autodidactic disabled (no learning overhead)
  • • Domain Extension enabled
  • • Explanation level 2-3

Maximum Capability (research/enterprise preset)

Full capabilities, comprehensive explanations:

  • • All Tetrad pillars enabled
  • • Maximum agents, full learning
  • • Explanation level 4-5
  • • All domain libraries active