TetraState：面向学习型世界模型的回归测试协议与测试框架版本化

View PDF HTML (experimental)

Abstract:Learned world models ship as software components whose model versions, preprocessing, and test harnesses change between releases. Work on model updates already counts per-sample regressions, or negative flips, that an accuracy check hides. Two problems remain: a test can pass under one configuration and fail under another, and a harness fault can produce erroneous scores and wrong verdicts without a crash. We present TetraState, a regression-testing protocol for paired possible/impossible video tests. Each test runs under four fixed model-native configurations; its number of correct verdicts assigns one of four outcome classes. Inconclusive tests and failed executions are reported separately alongside the verdict rate. Revisions are compared by per-test outcome transitions, and the harness is versioned as part of the system under test. On all 506 IntPhys 2 Main tests and two heterogeneous systems, Cosmos 3 Nano and V-JEPA 2, two V-JEPA 2 harness revisions move modal accuracy by only 0.2 and 2.6 points yet change 11 and 214 test labels; in the second, accuracy rises although fewer tests are judged correctly, and 43 regressions keep the modal verdict, so a modal-verdict flip check misses them. These results outline a research agenda on configuration-level regression oracles for learned components.

Subjects:	Software Engineering (cs.SE); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
Cite as:	arXiv:2610.41130 [cs.SE]
	(or arXiv:2610.41130v1 [cs.SE] for this version)
	https://doi.org/10.48550/arXiv.2610.41130 arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Junyu Huo [view email]
[v1] Fri, 30 Oct 2026 01:36:37 UTC (32 KB)