TRUST-REGION SALIENCY-GUIDED LOCAL SEARCH FOR INTERPRETABLE SEQUENCE DESIGN AT FIXED EDIT BUDGETS

Sara E. Pour

Confirmation: I have read and agree with the workshop's policy on behalf of myself and my co-authors.

Track: tiny / short paper (2-4 pages excluding references; extended abstract format)

Keywords: sequence design, discrete optimization, edit budget, trust region, saliency-guided search, Integrated Gradients, DeepSHAP, interpretability, BPNet, regulatory genomics, motif analysis, local search

TL;DR: SAGE-TRSwap designs DNA sequences under a fixed edit budget by using saliency-guided trust-region + swap moves to match target predictions while producing far more localized, interpretable edits.

Abstract: Discrete sequence design under a fixed edit budget can match target model outputs, but often returns dispersed, multi-cluster edits that are hard to interpret. We present $\textbf{SAGE-TRSwap}$, a saliency-guided trust-region local search that optimizes the same prediction loss as a Ledidi-style relaxation+pruning baseline while biasing proposals toward high-attribution regions and enabling budget-preserving swap refinements. Across 12 regulatory targets/tracks and 5 random starts per target (60 runs per budget), SAGE-TRSwap sharply reduces edit span and cluster count at all budgets while maintaining or improving absolute error.

Anonymization: This submission has been anonymized for double-blind review via the removal of identifying information such as names, affiliations, and identifying URLs.

Presenter: ~Sara_E._Pour1

Format: Yes, the presenting author will attend in person if this work is accepted to the workshop.

Funding: No, the presenting author of this submission does *not* fall under ICLR’s funding aims, or has sufficient alternate funding.

Submission Number: 71