Feasible-First Exploration for Constrained ML Deployment Optimization in Crash-Prone Hierarchical Search Spaces
Christian Lysenstøen · 2026
Read Feasible-First Exploration for Constrained ML Deployment Optimization in Crash-Prone Hierarchical Search Spaces by Christian Lysenstøen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-25073.txt.
Description
Lysenstøen studies feasible-first search for constrained deployment tuning in crash spaces. The space holds many invalid setups: tests may crash, exceed memory or violate latency limits. It proposes Thermal Budget Annealing, a feasible method mapping valid regions before warm-starting a Parzen Estimator. Two robustness tools handle hardware: timeouts aborting infeasible tests early. The authors present DeployBench, a benchmark suite with crash zones, hard constraints and unequal costs. On synthetic tests and real GPU deployment with five vision models across five targets, the hybrid improves discovery under constraints.
Catalog record
| CREATOR | Christian Lysenstøen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 54114 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.25073