Argument Evaluation in the Context of Assurance Case Confidence Modeling
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
In recent years, assurance cases have been gaining popularity across various domains, such as the railway, aeronautics, automotive and medical domains, as an important tool in the establishment of system safety. The assurance case is essentially an argument for the existence of a certain system property. The confidence that we may place in the validity of any such argument plays an important role in the decision-making process, both for the developer and the regulator. However, even though there is increasing interest in this research topic, it seems that there is no consensus on what the precise definition of assurance case confidence is, and therefore the approaches for its modeling and measurement vary. The concept of an assurance case argument is based on the ideas presented by Toulmin in his groundbreaking work [1]. He outlined a scheme for the layout of arguments, but did not provide guidelines for formal argument evaluation. Here we look into some works extending his ideas to incorporate a theory of argument evaluation, and offer our insights on what the implications are for the definition of confidence, as well as an approach that would prove suitable for its modeling. In essence, when we reason about the confidence one might place in an argument, we are trying to establish how well the argument corresponds to the notions of a 'good argument', as well as taking into account any and all sources of uncertainty that are inherent when we are faced with imperfect information. Even so, what we ultimately measure is not how true the conclusions of the argument are, but instead, how justifiable they are given our current knowledge.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 it