Research map / Learning theory and algorithms / Submodular and game algorithms
Recruitment Strategies That Take a Chance
Gregory Kehne, Ariel D. Procaccia, Jingyan Wang
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 PGreedy (sort by acceptance probability) instead of value-aware heuristics, penalty-only objective | PGreedy · penalty-only objective, convex loss | penalty V(S) | Improved |
| apply value-ignoring PGreedy to the full reward-minus-penalty objective with L2 loss | PGreedy · L2 loss, general objective | utility U(S) | Hurt |
| use XGreedy or XPGreedy under L2 loss | greedy heuristics · L2 loss, general objective | utility U(S) | Hurt |
| use the FPTAS for L2 loss instead of greedy heuristics | Algorithm 4 · L2 loss | additive approximation to U(S*) | Improved |
| use XPGreedy with L1+ loss in the low-value regime | XPGreedy · L1+ loss, x_max <= (1-c)lambda | approximation ratio | Mixed |
| use XGreedy with L1+ loss in the low-value regime | XGreedy · L1+ loss, x_max <= (1-c)lambda | approximation ratio | Mixed |
| replace greedy heuristics by OneSidedL1+ with value grouping | OneSidedL1+ · L1+ loss | approximation ratio | Improved |
| positive correlation between value and acceptance probability in synthetic data | XGreedy, XPGreedy, LowValueL1+ · synthetic instances, n=50, pmin=0.01 | objective | No clear effect |
| switch to no or negative correlation between value and probability | XGreedy vs XPGreedy · synthetic instances, n=50 | objective | Mixed |
| use LowValueL1+ instead of single greedy heuristic under negative correlation | LowValueL1+ · negative correlation, lambda in {1.5,5,30} | objective | Mixed |
| round probabilities up by constant factor in LowValueL1+ | LowValueL1+ · negative correlation, lambda=1.5 | objective | Hurt |
| use smaller multiplicative bucket sizes in rounding for LowValueL1+ | LowValueL1+ · clustered probabilities | objective | Improved |
Closest papers by meaning
- Addressing Bias in Online Selection with Limited Budget of Comparisons (NeurIPS 2024 · similarity 0.84)
- On Mitigating Affinity Bias through Bandits with Evolving Biased Feedback (ICML 2025 · similarity 0.83)
- A Principled Approach to Randomized Selection under Uncertainty: Applications to Peer Review and Grant Funding (NeurIPS 2025 · similarity 0.83)
- Matchings Under Biased and Correlated Evaluations (NeurIPS 2025 · similarity 0.82)
- Maximizing the Success Probability of Policy Allocations in Online Systems (AAAI 2024 · similarity 0.81)
- Centralized Selection with Preferences in the Presence of Biases (ICML 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.
