Computational Argumentation and Automatic Rule-Generation for Explainable Data-Driven Modeling
Bibliographic record
Abstract
The creation of data-driven models for classification problems requires increasing transparency and inferential explainability, especially in high-stakes domains such as health-care, finance, and policy making. Rule-based systems are widely regarded as a strong candidate for the development of models that are also comprehensible to humans. However, the generated rules are often considered individually with minimal or no consideration of their interactions. This research focuses on the adoption of computational argumentation techniques, which allow for rule-interaction for enhanced explainability. In other words, rules can be revoked when new information is introduced, essentially achieving the notion of non-monotonicity. In detail, an empirical work was designed to automatically extract inference rules from datasets of various multi-class classification tasks by using the Logic Learning Machine (LLM) approach. In turn, these rules were integrated within a structured argumentation framework, able to employ abstract argumentation semantics for conflict resolution among contradicting inferences. Findings demonstrated that the LLM technique can indeed extract compact rules with varying degrees of interpretability and predictive power. Furthermore, the argument-based models built on these rules demonstrated improved inferential and explanatory performance on certain datasets. Examples show how a Cohen’s kappa coefficient improved from 0.85 to 0.99 when applying the argumentation-based conflict resolution strategy to the same set of rules generated by LLM. The contribution to the body of knowledge offered to the community is both a customisable approach for rule-extraction from datasets for multi-class problems, via hyperparameter tuning, and a transparent integration strategy with computational argumentation, which is able to enhance human understanding and support justifiability.
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 | 0.000 | 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 | 0.001 | 0.003 |
| 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 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".