Tunable Domain Adaptation Using Unfolding

TL;DR

This paper introduces two interpretable unrolled network-based methods for tunable domain adaptation in regression tasks, enabling flexible adaptation to varying data distributions with improved performance.

Contribution

It proposes novel, interpretable unrolled network architectures for domain adaptation that can be tuned during inference, either via known domain parameters or data-driven inference.

Findings

Validated on compressed sensing problems with noise adaptation, gain calibration, and phase retrieval.

Achieved performance comparable or better than domain-specific models.

Outperformed traditional joint training baselines.

Abstract

Machine learning models often struggle to generalize across domains with varying data distributions, such as differing noise levels, leading to degraded performance. Traditional strategies like personalized training, which trains separate models per domain, and joint training, which uses a single model for all domains, have significant limitations in flexibility and effectiveness. To address this, we propose two novel domain adaptation methods for regression tasks based on interpretable unrolled networks--deep architectures inspired by iterative optimization algorithms. These models leverage the functional dependence of select tunable parameters on domain variables, enabling controlled adaptation during inference. Our methods include Parametric Tunable-Domain Adaptation (P-TDA), which uses known domain parameters for dynamic tuning, and Data-Driven Tunable-Domain Adaptation (DD-TDA), which infers domain adaptation directly from input data. We validate our approach on compressed sensing problems involving noise-adaptive sparse signal recovery, domain-adaptive gain calibration, and domain-adaptive phase retrieval, demonstrating improved or comparable performance to domain-specific models while surpassing joint training baselines. This work highlights the potential of unrolled networks for effective, interpretable domain adaptation in regression settings.