Research map / Video / Video segmentation and tracking
Voice for the Voiceless: Active Sampling to Detect Comments Supporting the Rohingyas
Shriphani Palakodety, Ashiqur R. KhudaBukhsh, Jaime G. Carbonell, Shriphani Palakodety, Ashiqur R. KhudaBukhsh, Jaime G. Carbonell
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 |
|---|---|---|---|
| Add comment embeddings as features to the n-gram SVM (ablation) | Voice-for-the-voiceless classifier · Voice-for-the-voiceless test set | F1 score: 76.34 ± 2.77% -> 78.28 ± 2.71 | Improved |
| Add comment embeddings as features to the n-gram SVM (ablation) | Voice-for-the-voiceless classifier · Voice-for-the-voiceless test set | AUC: 83.67 ± 2.61% -> 85.91 ± 2.32 | Improved |
| Add minority-class certainty sampling after comment nearest-neighbor sampling (ablation) | Voice-for-the-voiceless classifier · Voice-for-the-voiceless test set | F1 score: 43.02 ± 7.90% -> 71.75 ± 4.32% | Improved |
| Add minority-class certainty sampling after comment nearest-neighbor sampling (ablation) | Voice-for-the-voiceless classifier · Voice-for-the-voiceless test set | Accuracy: 82.04 ± 2.34% -> 75.95 ± 3.10% | Hurt |
| Add uncertainty sampling after certainty sampling (ablation) | Voice-for-the-voiceless classifier · Voice-for-the-voiceless test set | F1 score: 71.75 ± 4.32% -> 76.34 ± 2.77% | Improved |
| Add uncertainty sampling after certainty sampling (ablation) | Voice-for-the-voiceless classifier · Voice-for-the-voiceless test set | Accuracy: 75.95 ± 3.10% -> 75.38 ± 2.76% | Hurt |
| Use comment nearest-neighbor sampling instead of random sampling to find positive examples | Active sampling · Comments sampled for labeling | Positive comments: 10.66% -> 33.67% | Improved |
| Sample comments from nearest neighbors of high-confidence positive users | User-focused active sampling · Comments sampled for labeling | Positive comments: 10.66% -> 35% | Improved |
Closest papers by meaning
- A Multi-Task Benchmark for Abusive Language Detection in Low-Resource Settings (NeurIPS 2025 · similarity 0.84)
- Multi-Modal Discussion Transformer: Integrating Text, Images and Graph Transformers to Detect Hate Speech on Social Media (AAAI 2024 · similarity 0.83)
- Spanning the Spectrum of Hatred Detection: A Persian Multi-Label Hate Speech Dataset with Annotator Rationales (AAAI 2024 · similarity 0.83)
- TRACE: Textual Relevance Augmentation and Contextual Encoding for Multimodal Hate Detection (AAAI 2026 · similarity 0.82)
- Bridging the Data Provenance Gap Across Text, Speech, and Video (ICLR 2025 · similarity 0.81)
- IndicCONAN: A Multilingual Dataset for Combating Hate Speech in Indian Context (AAAI 2024 · similarity 0.81)
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.
