Beyond Black-Box Labels: Interpretable Criteria for Diagnosing Subjective NLP Tasks
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911634779275264 |
|---|---|
| author | Rair, Nisrine Goupil, Alban Vrabie, Valeriu Chochoy, Emmanuel |
| author_facet | Rair, Nisrine Goupil, Alban Vrabie, Valeriu Chochoy, Emmanuel |
| contents | Subjective NLP datasets typically aggregate annotator judgments into a single gold label, making it difficult to diagnose whether disagreement reflects unclear criteria, collapsed distinctions, or legitimate plurality. We propose a \emph{schema-level diagnostic} for auditing expert-designed annotation schemas \emph{prior to} gold-label commitment, using only multi-annotator criterion judgments. The diagnostic separates two failure modes: unstable criteria with hard-to-operationalize boundaries, and systematic overlap that blurs the boundaries between mutually exclusive categories. Applied to persuasive value extraction in commercial documents, we find that disagreement is not diffuse: instability concentrates in a few criteria, while nearly half of covered sentences activate multiple categories. These signals align with where domain experts disagree, yielding an evidence-based audit for tightening guidelines, revising category structure, or reconsidering the annotation paradigm. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_17022 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Beyond Black-Box Labels: Interpretable Criteria for Diagnosing Subjective NLP Tasks Rair, Nisrine Goupil, Alban Vrabie, Valeriu Chochoy, Emmanuel Computation and Language Artificial Intelligence I.2.7 Subjective NLP datasets typically aggregate annotator judgments into a single gold label, making it difficult to diagnose whether disagreement reflects unclear criteria, collapsed distinctions, or legitimate plurality. We propose a \emph{schema-level diagnostic} for auditing expert-designed annotation schemas \emph{prior to} gold-label commitment, using only multi-annotator criterion judgments. The diagnostic separates two failure modes: unstable criteria with hard-to-operationalize boundaries, and systematic overlap that blurs the boundaries between mutually exclusive categories. Applied to persuasive value extraction in commercial documents, we find that disagreement is not diffuse: instability concentrates in a few criteria, while nearly half of covered sentences activate multiple categories. These signals align with where domain experts disagree, yielding an evidence-based audit for tightening guidelines, revising category structure, or reconsidering the annotation paradigm. |
| title | Beyond Black-Box Labels: Interpretable Criteria for Diagnosing Subjective NLP Tasks |
| topic | Computation and Language Artificial Intelligence I.2.7 |
| url | https://arxiv.org/abs/2604.17022 |