Research map / Label-efficient and robust learning / Recognition and long-tail
AdaFlock: Adaptive Feature Discovery for Human-in-the-loop Predictive Modeling
Ryusuke Takahama, Yukino Baba, Nobuyuki Shimizu, Sumio Fujita, Hisashi Kashima
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 |
|---|---|---|---|
| adaptive example selection for feature definition (AdaFlock vs StandardFlock) (ablation) | feature crowdsourcing · Paintings trial 1 | testing accuracy: 0.755 -> 0.820 | Improved |
| adaptive example selection for feature definition (AdaFlock vs StandardFlock) (ablation) | feature crowdsourcing · Paintings trial 2 | testing accuracy: 0.745 -> 0.705 | Hurt |
| adaptive example selection (AdaFlock vs StandardFlock) (ablation) | feature crowdsourcing · Smiles trial 1 | testing accuracy: 0.715 -> 0.780 | Improved |
| adaptive example selection (AdaFlock vs StandardFlock) (ablation) | feature crowdsourcing · Smiles trial 2 | testing accuracy: 0.720 -> 0.740 | Improved |
| AdaFlock vs crowd prediction | classification · Paintings trial 1 | testing accuracy: 0.560 -> 0.820 | Improved |
| AdaFlock vs crowd prediction | classification · Paintings trial 2 | testing accuracy: 0.580 -> 0.705 | Improved |
| AdaFlock vs crowd prediction | classification · Smiles trial 1 | testing accuracy: 0.720 -> 0.780 | Improved |
| AdaFlock vs crowd prediction | classification · Smiles trial 2 | testing accuracy: 0.715 -> 0.740 | Improved |
| AdaFlock crowd features vs off-the-shelf Inception-v3 features | classification · Paintings trial 1 | testing accuracy: 0.755 -> 0.820 | Improved |
| AdaFlock crowd features vs off-the-shelf Inception-v3 features | classification · Paintings trial 2 | testing accuracy: 0.735 -> 0.705 | Hurt |
| increase AdaFlock iterations (feature discovery of ground feature at iteration 5) | AdaFlock · Paintings trial 1 | testing accuracy: 10% | Improved |
| increase AdaFlock iterations (discovery of short-duration feature) | AdaFlock · Smiles | testing accuracy: 8% | Improved |
Closest papers by meaning
- KFNN: K-Free Nearest Neighbor For Crowdsourcing (NeurIPS 2024 · similarity 0.83)
- Lean Multiclass Crowdsourcing (CVPR 2018 · similarity 0.83)
- Making Better Use of the Crowd: How Crowdsourcing Can Advance Machine Learning Research (JMLR 2018 · similarity 0.83)
- Max-MIG: an Information Theoretic Approach for Joint Learning from Crowds (ICLR 2019 · similarity 0.83)
- Human-in-the-Loop Feature Selection (AAAI 2019 · similarity 0.83)
- Crowdsourcing via Pairwise Co-occurrences: Identifiability and Algorithms (NeurIPS 2019 · similarity 0.83)
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.
