Research map / Label-efficient and robust learning / OOD and anomaly detection
A Human-AI Loop Approach for Joint Keyword Discovery and Expectation Estimation in Micropost Event Detection
Akansha Bhardwaj, Jie Yang, Philippe Cudré-Mauroux
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 iterative human-AI keyword discovery and expectation estimation instead of the single-keyword baseline | LR event detection · CyberAttack and PoliticianDeath | AUC: 18.38% | Improved |
| Use iterative human-AI keyword discovery and expectation estimation instead of the single-keyword baseline | MLP event detection · CyberAttack and PoliticianDeath | AUC: 30.27% | Improved |
| Discover keywords using model-crowd disagreement rather than query expansion | MLP event detection · CyberAttack | AUC: 4.62% | Improved |
| Discover keywords using model-crowd disagreement rather than query expansion | MLP event detection · PoliticianDeath | AUC: 52.58% | Improved |
| Use the crowdsourcing budget for keyword discovery instead of obtaining additional training labels | Event detection · CyberAttack | AUC: 15.23% vs 0.87% | Improved |
| Use the crowdsourcing budget for keyword discovery instead of obtaining additional training labels | Event detection · PoliticianDeath | AUC: 33.42% vs 1.06% | Improved |
| Infer keyword expectation jointly using crowd labels and model predictions rather than majority voting | MLP event detection · CyberAttack | AUC: 0.4% | Improved |
| Infer keyword expectation jointly using crowd labels and model predictions rather than majority voting | MLP event detection · PoliticianDeath | AUC: 1.19% | Improved |
| Continue discovering keywords after the model has largely converged | Human-AI loop event detection · CyberAttack and PoliticianDeath | AUC | Hurt |
Closest papers by meaning
- Position: On the Possibilities of AI-Generated Text Detection (ICML 2024 · similarity 0.82)
- Human Texts Are Outliers: Detecting LLM-generated Texts via Out-of-distribution Detection (NeurIPS 2025 · similarity 0.82)
- Hierarchical and Incremental Structural Entropy Minimization for Unsupervised Social Event Detection (AAAI 2024 · similarity 0.81)
- Training-free LLM-generated Text Detection by Mining Token Probability Sequences (ICLR 2025 · similarity 0.81)
- Multiscale Positive-Unlabeled Detection of AI-Generated Texts (ICLR 2024 · similarity 0.80)
- Online Detection of LLM-Generated Texts via Sequential Hypothesis Testing by Betting (ICML 2025 · similarity 0.80)
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.
