Can Large Language Models Assist with SOTIF Scenario Generation?
Bibliographic record
Abstract
Combinations of operating conditions can trigger a system to behave in a hazardous manner, even in the absence of malfunction. ISO 21448 - Road vehicles - Safety of the intended functionality (referred to as “SOTIF”) describes strategies for managing this type of risk in automotive systems. One strategy includes the identification of operating scenarios that might lead to the occurrence of a hazard. Crafting scenarios is a technically challenging and labor-intensive task that requires sustained creative engagement, and the consequence of inadequate SOTIF analyses can be severe. This paper introduces Heraclitus, an engineering method and prototype software tool for performing SOTIF scenario generation with the support of a large language model. Large language models are a novel type of generative artificial intelligence targeted at natural language processing and generation that exhibit remarkable performance in a range of natural language applications that have historically been difficult for conventional artificial intelligence systems. As such, there is an opportunity to use these models, in collaboration with humans, to support SOTIF scenario creation. The goal of Heraclitus is to allow analysts to rapidly produce a comprehensive set of SOTIF scenarios that can be used as the basis for on-going SOTIF risk management. A preliminary control trial of Heraclitus was conducted, in which six system safety experts were asked to create SOTIF scenarios with and without the support of a large language model. Results indicate that these models show promise in supporting SOTIF analysis and are capable of generating useful SOTIF scenarios.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.037 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.007 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.006 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".