Budgeted stream-based active learning via adaptive submodular maximization
Authors: Kaito Fujii, Hisashi Kashima
NeurIPS 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In addition we empirically demonstrate their effectiveness by comparing with existing heuristics on common benchmark datasets. |
| Researcher Affiliation | Academia | Kaito Fujii Kyoto University JST, ERATO, Kawarabayashi Large Graph Project fujii@ml.ist.i.kyoto-u.ac.jp Hisashi Kashima Kyoto University kashima@i.kyoto-u.ac.jp |
| Pseudocode | Yes | Algorithm 1 Adaptive Stream algorithm & Adaptive Secretary algorithm |
| Open Source Code | No | The paper does not explicitly state that source code for the described methodology is released or provide a link to a code repository. |
| Open Datasets | Yes | We conducted experiments on two benchmark datasets, WDBC3 and MNIST4. 3https://archive.ics.uci.edu/ml/datasets/Breast+Cancer+Wisconsin+(Diagnostic) 4http://yann.lecun.com/exdb/mnist/ |
| Dataset Splits | No | The paper mentions initial training and selecting instances but does not provide explicit training, validation, or test dataset split percentages or exact counts for the entire dataset used in the experiments. |
| Hardware Specification | No | The paper does not provide specific details about the hardware (e.g., CPU, GPU models, memory) used to run the experiments. |
| Software Dependencies | No | The paper mentions using 'linear SVM' but does not specify any software dependencies with version numbers (e.g., Python, PyTorch, scikit-learn versions). |
| Experiment Setup | Yes | The number of hypotheses sampled at each time is set N = 1000 in all settings. In every setting, we first randomly select 10 instances for the initial training of a classifier and after that, select k 10 instances with each method. We use the linear SVM trained with instances labeled so far to judge the uncertainty. We conducted experiments on two benchmark datasets, WDBC3 and MNIST4...We evaluate the performance through 100 trials, where at each time an order in which the instances arrive is generated randomly. |