RASP-Tuner: Retrieval-Augmented Soft Prompts for Context-Aware Black-Box Optimization in Non-Stationary Environments

TL;DR

RASP-Tuner is a novel context-aware black-box optimizer that leverages retrieval-augmented soft prompts to adapt efficiently in non-stationary environments, outperforming traditional methods in various benchmarks.

Contribution

It introduces a retrieval-based soft prompt framework for online tuning, combining regime identification with low-dimensional adaptation to improve efficiency and performance.

Findings

RASP-Tuner matches or exceeds regret of GP-UCB and CMA-ES on synthetic benchmarks.

It achieves 8-12 times lower wall-clock time per step compared to sliding-window GP-UCB.

The method performs well on real-world streaming data, demonstrating practical applicability.

Abstract

Many deployed systems expose black-box objectives whose minimizing configuration shifts with an externally observed context. When contexts revisit a small set of latent regimes, an optimizer that discards history pays repeated adaptation cost; when each step must remain inexpensive, full Gaussian-process (GP) refits at high observation counts are difficult to sustain. We cast online tuning as context-conditioned regret minimization and present RASP-Tuner, which instantiates a decomposition motivated by first principles: (i) identify a regime proxy by retrieving similar past contexts; (ii) predict short-horizon loss with a mixture-of-experts surrogate whose input concatenates parameters, context, and a retrieved soft prompt; (iii) adapt chiefly in a low-dimensional prompt subspace, invoking full surrogate updates only when scalarized error or disagreement spikes. A RealErrorComposer maps heterogeneous streaming metrics to [0,1] via EMA-stabilized logistic scores, supplying a single differentiable training target. On nine synthetic non-stationary benchmarks, an adversarial-context sanity check, and three tabular real-world streams (Section on real-world experiments), RASP-Tuner improves or matches cumulative regret relative to our GP-UCB and CMA-ES implementations on seven of nine synthetic tasks under paired tests at horizon T=100, while recording 8-12 times lower wall-clock per step than sliding-window GP-UCB on identical hardware. Idealized analysis in a cluster-separated, strongly convex regime model (RA-GD) supplies sufficient conditions for bounded dynamic regret; the deployed pipeline violates several of these premises, and we articulate which gaps remain open.