Research map / Label-efficient and robust learning / Recognition and long-tail
Bias Reduction via End-to-End Shift Learning: Application to Citizen Science
Di Chen, Carla P. Gomes
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 |
|---|---|---|---|
| apply covariate-shift correction with SCN rather than train without bias correction | multi-species distribution modeling · eBird, NLCD features | AUC: 77.86 -> 80.34 | Improved |
| apply covariate-shift correction with SCN rather than train without bias correction | multi-species distribution modeling · eBird, Google Earth images | AUC: 80.93 -> 83.80 | Improved |
| use end-to-end SCN instead of the strongest listed baseline, DFW | multi-species distribution modeling · eBird, NLCD features | AUC: 79.38 -> 80.34 | Improved |
| use end-to-end SCN instead of the strongest listed baseline, DFW | multi-species distribution modeling · eBird, Google Earth images | AUC: 81.99 -> 83.80 | Improved |
| add feature space mean matching loss (ablation) | SCN · eBird, NLCD features | AUC: 79.53 -> 80.34 | Improved |
| add feature space mean matching loss (ablation) | SCN · eBird, Google Earth images | AUC: 82.35 -> 83.80 | Improved |
| add discriminative loss (ablation) | SCN · eBird, NLCD features | AUC: 79.58 -> 80.34 | Improved |
| add discriminative loss (ablation) | SCN · eBird, Google Earth images | AUC: 82.49 -> 83.80 | Improved |
| use moving-average rather than batch-wise estimation for FSMM (ablation) | SCN · eBird, NLCD features | AUC: 80.09 -> 80.34 | Improved |
| use moving-average rather than batch-wise estimation for FSMM (ablation) | SCN · eBird, Google Earth images | AUC: 83.44 -> 83.80 | Improved |
| use FSMM alone to learn the shift factor, without the discriminative loss (ablation) | SCN · eBird multi-species modeling | predictive performance | Mixed |
Closest papers by meaning
- WILDS: A Benchmark of in-the-Wild Distribution Shifts (ICML 2021 · similarity 0.84)
- BREEDS: Benchmarks for Subpopulation Shift (ICLR 2021 · similarity 0.83)
- An Information-theoretic Approach to Distribution Shifts (NeurIPS 2021 · similarity 0.82)
- DivShift: Exploring Domain-Specific Distribution Shifts in Large-Scale, Volunteer-Collected Biodiversity Datasets (AAAI 2025 · similarity 0.81)
- DeNetDM: Debiasing by Network Depth Modulation (NeurIPS 2024 · similarity 0.80)
- SEBRA : Debiasing through Self-Guided Bias Ranking (ICLR 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.
