A Modularized Framework for Explaining Black Box Classifiers for Text Data
Bibliographic record
Abstract
The cumbersome amount of textual data produced in social media and in the new digital life makes the usage of automatic decision systems necessary for acting on text. The most widely adopted natural language processing approaches guarantee high accuracy but are black-box systems, that hide the logic of their internal decision processes. Since in various applications there is the need to unveil the reasons for the classification of different texts, the urge to explain black-box behaviour is growing among scientists. Thus, we propose a local model-agnostic method for interpreting text classifiers. Our method explains the decision of a text classifier on a given document by generating similar samples in its vicinity. The new samples are generated by replacing words of the document under analysis with their synonyms, antonyms, hyponyms, hypernyms, and definitions. Finally, these synthetic texts are used to train a decision tree that enables the user to identify important words explaining the classification outcome. An inspection of the synthetic documents generated by our proposal together with a set of words appropriately highlighted explain why the black box assigns a certain label to a given document. Deep and wide experimentation on various datasets and classifiers shows the effectiveness of our proposal and that its performance overcomes state-of-the-art methods.
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 | 0.001 | 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 | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.002 | 0.002 |
| 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".