Research map / Learning theory and algorithms / Fairness
Argumentative Debates for Transparent Bias Detection
Hamed Ayoobi, Nico Potyka, Anna Rapberger, Francesca Toni
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use ABIDE instead of IRB | bias detector · Synthetic classifier G2 (K=50) | F1-score: 0.00 -> 0.98 | Improved |
| use ABIDE instead of IRB | bias detector · Synthetic classifier L1 (K=200) | F1-score: 0.36 -> 1.00 | Improved |
| use ABIDE instead of IRB | bias detector · Synthetic classifier L1 (K=50) | F1-score: 0.61 -> 1.00 | Improved |
| use ABIDE instead of IRB | bias detector · Synthetic classifier G1 (K=50) | F1-score: 0.95 -> 0.95 | No clear effect |
| use ABIDE instead of IRB | bias detector · Synthetic classifier G1 (K=50) | runtime (s): 28.22 -> 3.87 | Improved |
| use ABIDE instead of IRB | bias detector · Synthetic classifier L1 (K=200) | runtime (s): 97.16 -> 5.07 | Improved |
| combine multiple neighbourhoods (K=50,100,200) instead of the largest alone, ABIDE vs IRB | bias detector · Synthetic classifier L1 | F1-score: 0.36 -> 1.00 | Improved |
| use ABIDE instead of IRB | bias detector · COMPAS logistic regression, African-American | biased cases found: 2 -> 77 | Improved |
| use ABIDE instead of IRB | bias detector · COMPAS logistic regression, Hispanic | biased cases found: 26 -> 13 | Mixed |
| use ABIDE instead of IRB | bias detector · Bank Marketing logistic regression, YoungOrOld | biased cases found: 20 -> 78 | Improved |
| use ABIDE instead of IRB | bias detector · ChatGPT-4o on COMPAS, African-American | biased cases found: 0 -> 129 | Improved |
| use ABIDE instead of IRB | bias detector · ChatGPT-4o on Bank Marketing, Single | biased cases found: 0 -> 18 | Improved |
Closest papers by meaning
- Identifying Reasons for Bias: An Argumentation-Based Approach (AAAI 2024 · similarity 0.89)
- Collaborative Disagreement Resolution for Scalable Oversight (ICML 2026 · similarity 0.88)
- When Debate Helps: Proposal Supply and Verification-Aware Readout in Multi-Agent Reasoning (NeurIPS 2026 · similarity 0.86)
- OpenDebateEvidence: A Massive-Scale Argument Mining and Summarization Dataset (NeurIPS 2024 · similarity 0.85)
- On scalable oversight with weak LLMs judging strong LLMs (NeurIPS 2024 · similarity 0.85)
- Multi-Agent Debate for LLM Judges with Adaptive Stability Detection (NeurIPS 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
