Research map / Learning theory and algorithms / Sample complexity
Crowdsourcing with Arbitrary Adversaries
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 colluding corrupted workers up to 23 of 50 | Alg. 1 · Synthetic, n=50, m=5000 | prediction and estimation error | No clear effect |
| add six or more corrupted workers | literature methods (KOS, GKM, RoE, EoR, S-EM, TE) · Synthetic | prediction error | Hurt |
| use linear weights (Alt-Alg. 1) instead of logarithmic weights (Alg. 1) (ablation) | Algorithm 1 · Synthetic | prediction error | Mixed |
| increase number of tasks m from 5000 to 20000 | Alg. 1 / Alt-Alg. 1 · Synthetic, 23 corrupted | prediction error | Helped |
| increase number of tasks m from 5000 to 20000 | Alg. 1 · Synthetic, 23 corrupted | estimation error | Helped |
| prefilter corrupted workers with soft-penalty algorithm of Jagabathula et al. (2017) | baseline combination · Synthetic | corrupted worker identification | Hurt |
| compare Alt-Alg. 1 with S-EM10 | Alt-Alg. 1 · Real data (six sets) | prediction error | Mixed |
| Alg. 1 vs Alt-Alg. 1 (ablation) | Algorithm 1 · Real data | prediction error | Hurt |
| workers deviate from one-coin model | Algorithm 1 · Bird | prediction error | Hurt |
| corrupt random workers in real data | Alt-Alg. 1 · Real data (Bird, Dog, Duchenne) | prediction error | Helped |
Closest papers by meaning
- Avoiding Imposters and Delinquents: Adversarial Crowdsourcing and Peer Prediction (NeurIPS 2016 · similarity 0.87)
- A Minimax Optimal Algorithm for Crowdsourcing (NeurIPS 2017 · similarity 0.86)
- Analysis of Minimax Error Rate for Crowdsourcing and Its Application to Worker Clustering Model (ICML 2018 · similarity 0.86)
- Exact Exponent in Optimal Rates for Crowdsourcing (ICML 2016 · similarity 0.85)
- Evaluating LLM-contaminated Crowdsourcing Data Without Ground Truth (NeurIPS 2025 · similarity 0.85)
- Adversarial Crowdsourcing Through Robust Rank-One Matrix Completion (NeurIPS 2020 · similarity 0.85)
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.