Compositional Dataflow Analysis via Abstract Transition Systems
Bibliographic record
Abstract
<div> Fully precise static analysis of program behavior is undecidable in theory and infeasible in practice. While individual static analyses can prove specialized invariants about program behavior or resolve narrow ambiguities, real programs are complex composites of many different abstractions and programming patterns. State-of-the-art approaches to program analysis thus generally rely on composite analyses, which combine multiple distinct program analyses, each focused on specific aspects of program behavior. However, the diversity of static analysis techniques, implementation strategies, information representations, and the variation in the forms of abstraction that each analysis requires and produces makes analysis cofmposition a challenging task. In practice analysis developers either customize the coupling of pairs of analyses, where an analysis that wants information from analysis A develops logic to access A's API, or they define specific abstraction layers that enumerate and thus restrict the forms of information that the analyses may communicate. This paper proposes an approach to composing multiple static analyses that is applicable to any static analysis, enables arbitrary analyses to be composed, and facilitates information sharing independently of the analyses' internal abstractions or even their abstract domains. It introduces a portable interface based on abstract transition systems, an abstraction over execution traces, that may be used in the implementation of any analysis. This interface allows any client (a tool or another analysis) to query any analysis' results without any knowledge of the analysis' internal abstractions or implementation details, or restrictions on the categories of information that may be communicated. The interface supports the sequential composition of analysis passes, but also more 'tight' forms of composition in which multiple analyses cooperatively compute a mutual fixed point, without necessarily being aware of each other. Our overall approach (a) lowers the bar of entry for constructing * Uploaded to HAL on </div>
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 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 | 0.003 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, 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".