Adaptive Intelligence

Developer guide for configuring and using Adaptive Intelligence—self-improvement and continuous learning from optimization experience.

Overview

Adaptive Intelligence enables the system to learn from every optimization, recognize similar problems, and improve performance over time. It uses problem signature detection, pattern recognition, and cross-problem learning to continuously enhance optimization strategies.

How It Works

1.Problem Signature Detection: System recognizes similar problems based on characteristics, constraints, and patterns
2.Strategy Recall: Retrieves successful strategies from past optimizations of similar problems
3.Pattern Learning: Identifies patterns across different problem types to improve strategy selection
4.Continuous Improvement: Each optimization adds to the knowledge base, improving future performance

Learning Stores

Adaptive Intelligence uses two types of learning stores:

Private Learning Store

Your organization's private knowledge base. Stores patterns and strategies learned from your optimizations. Never shared with other organizations.

Public Learning Store

Shared knowledge base with anonymized patterns from all users. Provides broader learning but may not be suitable for sensitive problems.

Simple Configuration

Enable Adaptive Intelligence with a simple boolean flag:

Enable Adaptive Intelligence

from sematryx import Sematryx

client = Sematryx(api_key="sk-your-api-key")

# Enable Adaptive Intelligence
result = client.optimize(
    objective="minimize",
    variables=[{"name": "x", "bounds": (-5, 5)}, {"name": "y", "bounds": (-5, 5)}],
    objective_function=sphere,
    use_adaptive_intelligence=True
)

Learning Configuration

Control what the system learns from and writes to:

Configure Learning Behavior

from sematryx import Sematryx

client = Sematryx(api_key="sk-your-api-key")

# Configure learning behavior
result = client.optimize(
    objective="minimize",
    variables=[{"name": "x", "bounds": (-5, 5)}, {"name": "y", "bounds": (-5, 5)}],
    objective_function=sphere,
    learning={
        "read_from_private": True,   # Learn from your past optimizations
        "write_to_private": True,    # Store result for future learning
        "read_from_public": False    # Don't use public patterns (privacy)
    }
)

Learning Options

read_from_private (bool, default: False)
Learn from your organization's past optimizations. Recommended for recurring problems.
write_to_private (bool, default: False)
Store optimization results in your private learning store for future use.
read_from_public (bool, default: False)
Learn from anonymized public patterns. Useful for novel problems but may not be suitable for sensitive data.

Advanced Configuration

Fine-tune Adaptive Intelligence behavior with advanced options:

Advanced Adaptive Configuration

from sematryx import Sematryx

client = Sematryx(api_key="sk-your-api-key")

# Advanced Adaptive Intelligence configuration
result = client.optimize(
    objective="minimize",
    variables=[{"name": "x", "bounds": (-5, 5)}, {"name": "y", "bounds": (-5, 5)}],
    objective_function=sphere,
    intelligence_config={
        "use_adaptive_intelligence": True,
        "adaptive": {
            "learning_enabled": True,
            "cross_problem_learning": True,  # Learn across different problem types
            "memory_retention": 90,  # Days to retain learned patterns
            "meta_learning": True  # Enable meta-learning capabilities
        }
    }
)

Configuration Options

learning_enabled (bool, default: True)
Enable the learning system. Disable to prevent any learning operations.
cross_problem_learning (bool, default: True)
Allow learning across different problem types. Enables meta-learning capabilities.
memory_retention (int, default: 90)
Number of days to retain learned patterns. Older patterns are gradually deprecated.
meta_learning (bool, default: False)
Enable meta-learning to improve how the system learns. Advanced feature for long-term optimization projects.

REST API Configuration

Configure Adaptive Intelligence via REST API:

REST API - Adaptive Intelligence

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,
    "intelligence_config": {
      "use_adaptive_intelligence": true,
      "adaptive": {
        "learning_enabled": true,
        "cross_problem_learning": true,
        "memory_retention": 90,
        "meta_learning": true
      }
    },
    "learning": {
      "read_from_private": true,
      "write_to_private": true
    }
  }'

JavaScript SDK Configuration

Configure Adaptive Intelligence using the JavaScript SDK:

JavaScript SDK - Adaptive Intelligence

import { Sematryx } from '@sematryx/javascript-sdk'

const client = new Sematryx('sk-your-api-key')

// Enable Adaptive Intelligence
const result = await client.optimize({
  objective: 'minimize',
  variables: [
    { name: 'x', bounds: [-5, 5] },
    { name: 'y', bounds: [-5, 5] }
  ],
  objectiveFunction: sphere,
  learning: {
    readFromPrivate: true,
    writeToPrivate: true
  },
  intelligenceConfig: {
    adaptive: {
      learningEnabled: true,
      crossProblemLearning: true,
      memoryRetention: 90,
      metaLearning: true
    }
  }
})

Best Practices

•Enable for recurring problems: Adaptive Intelligence is most valuable when you run similar optimizations regularly.
•Use private learning for sensitive data: Keep read_from_public disabled when working with proprietary or sensitive problems.
•Write to private store: Enable write_to_private to build your organization's knowledge base over time.
•Adjust memory retention: Increase retention for stable problem types, decrease for rapidly changing environments.
•Enable meta-learning for long-term projects: Use meta_learning when running optimization projects over months or years.