CADAQUES

Cost-Aware Dual Architecture for QUery-Efficient diScovery

An open-source framework for autonomous discovery campaigns: any oracle, any driver, one budget. Every query counts.

CADAQUES decouples autonomous discovery into two symmetric protocols. A metered Oracle abstracts anything that answers queries at a price — a simulator, a laboratory instrument, an analytic function. A Driver abstracts anything that decides what to ask next — random search, Bayesian optimization, gradient methods, LLM agents. Between them sits the framework’s one structural commitment: cost is a first-class primitive. Every oracle query and every driver decision is priced in heterogeneous currencies and charged against a single campaign Budget.

Installation

pip install cadaques          # core: NumPy only
pip install "cadaques[bo]"    # + the Gaussian-process driver

Guide

• Quickstart
  • The campaign is a durable object (since 0.2)
  • From a spreadsheet to the next experiments (0.3)
• Concepts
  • The dual architecture
  • Declared vs. settled cost
  • Both sides are metered
  • Budget-aware strategies
  • Multi-currency budgets

Reference

• Core
  • Protocols
  • Cost and budget
  • Campaign
  • Ledger
  • Tasks and search spaces
  • Envelopes and records
  • Events
  • Campaign specifications
  • Artifacts
  • Resources
• Oracles
  • AnalyticOracle
  • quadratic_bowl()
  • T_C_EXACT
  • Ising2DOracle
• Drivers
  • RandomDriver
  • AnnealedLocalDriver
• Runtime
  • Campaign loop
  • Derived state
  • Replay, checkpoint, resume
  • Recommendation mode
• Statistics
  • Comparison
  • compare()
  • paired_bootstrap_ci()
  • paired_campaigns()
  • wilcoxon_signed_rank()
• Verticals
  • Chemistry
• Conformance suites
  • check_driver_contract()
  • check_oracle_contract()
  • check_resource_contract()

Citation

If you use CADAQUES in academic work, please cite it: doi:10.5281/zenodo.21293589 (concept DOI; per-release DOIs are listed in the CHANGELOG) and the paper: arXiv:2607.16127.