| 2019 |
A Robust Zero-Sum Game Framework for Pool-based Active Learning |
Zhu et al. |
AISTATS |
- |
|
distributationally, Classifier, None, Tra, Hard |
inefficiency, sampling bias, and sensitivity to imbalanced data distribu- tion. |
|
| 2019 |
HS2: Active learning over hypergraphs with pointwise and pairwise queries |
Chien et al. |
AISTATS |
- |
HAL problems |
pointwise queries, Hypergraph, None, Tra, Hard |
Hopkins 155 dataset |
|
| 2019 |
Region-Based Active Learning |
Cortes et al. |
AISTATS |
- |
Classification |
region-based AL, DNNs, None, Tra, Hard |
UCI:magic04, nomao, shuttle, a9a, ijcnn1, codrna, skin, covtype. |
We give a detailed theoretical analysis of ORIWAL, including gen- eralization error guarantees and bounds on the number of points labeled, in terms of both the hypothesis set used in each region and the prob- ability mass of that region. |
| 2017 |
Lower Bounds on Active Learning for Graphical Model Selection |
Scarlett and Cevher |
AISTATS |
- |
Graphical Model Selection |
mutual information, Graphical Model, None, Tra, Hard |
None |
|
| 2017 |
Near-optimal Bayesian Active Learning with Correlated and Noisy Tests |
Chen et al. |
AISTATS |
- |
how people make risky decisions, pairwise comparisons |
maximizes the gain in a surrogate objective, BNNs, None, Tra, hard |
MovieLens 100k dataset |
Our analysis relies on an information-theoretic auxiliary function to track the progress of ECED, and utilizes adaptive submodular- ity to attain the near-optimal bound. |