LLM-based pipeline generates diverse scenarios from NHTSA crash records for ADS testing in Metadrive simulator, identifying failures in limited tests.
Data -efficient learning via clustering -based sensitivity sampling: Foundation models and beyond
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
ASSS uses an adversarial selector and Gumbel-Softmax relaxation to retain 98.9% task performance with only 30% of the data by preferentially keeping boundary samples.
citing papers explorer
-
Scenario Generation for Testing of Autonomous Driving Systems Using Real-World Failure Records
LLM-based pipeline generates diverse scenarios from NHTSA crash records for ADS testing in Metadrive simulator, identifying failures in limited tests.
-
ASSS: A Differentiable Adversarial Framework for Task-Aware Data Reduction
ASSS uses an adversarial selector and Gumbel-Softmax relaxation to retain 98.9% task performance with only 30% of the data by preferentially keeping boundary samples.