Active Learning with Oracle Epiphany
Authors: Tzu-Kuo Huang, Lihong Li, Ara Vartanian, Saleema Amershi, Jerry Zhu
NeurIPS 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We present a theoretical analysis of active learning with more realistic interactions with human oracles...Our analysis shows that active learning is possible with oracle epiphany...This is verified with simulations in Section 5, which highlights the nuanced dependency between query complexity and epiphany parameters. |
| Researcher Affiliation | Collaboration | Tzu-Kuo Huang Uber Advanced Technologies Group Pittsburgh, PA 15201; Lihong Li Microsoft Research Redmond, WA 98052; Ara Vartanian University of Wisconsin Madison Madison, WI 53706; Saleema Amershi Microsoft Research Redmond, WA 98052; Xiaojin Zhu University of Wisconsin Madison Madison, WI 53706 |
| Pseudocode | Yes | Algorithm 1 EPICAL; Algorithm 2 Oracular-EPICAL |
| Open Source Code | No | The paper does not provide concrete access to source code for the methodology described. |
| Open Datasets | Yes | We consider the binary classification task of 5 vs. other digits on MNIST [Le Cun et al., 1998]. |
| Dataset Splits | No | The paper describes a training set of 60,000 examples and a testing set of 10,000 examples for the MNIST dataset but does not explicitly mention a separate validation split with specific counts or percentages. |
| Hardware Specification | No | The paper does not provide specific hardware details (exact GPU/CPU models, processor types, or memory amounts) used for running its experiments. |
| Software Dependencies | No | The paper mentions using 'online passive logistic regression' but does not provide specific software names with version numbers for replication. |
| Experiment Setup | No | The paper describes experimental scenarios such as varying the oracle epiphany parameter beta and different unknown regions (U), but does not specify concrete hyperparameter values or detailed training configurations (e.g., learning rate, batch size, number of epochs) for its models. |