Binary Classification from Positive Data with Skewed Confidence
Authors: Kazuhiko Shinoda, Hirotaka Kaji, Masashi Sugiyama
IJCAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We demonstrate the effectiveness of the proposed method through a synthetic experiment with simple linear models and benchmark problems with neural network models. We also apply our method to drivers drowsiness prediction to show that it works well with a real-world problem where confidence is obtained based on manual annotation. |
| Researcher Affiliation | Collaboration | Kazuhiko Shinoda1 , Hirotaka Kaji1 and Masashi Sugiyama2,3 1Frontier Research Center, Toyota Motor Corporation 2Center for Advanced Intelligence Project, RIKEN 3Graduate School of Frontier Sciences, The University of Tokyo |
| Pseudocode | No | The paper does not contain any clearly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper mentions implementing experiments using PyTorch and Scikit-learn and provides a link to PyTorch's website, but it does not state that the code for their own methodology is open-source or provide a link to it. |
| Open Datasets | Yes | The Fashion-MNIST dataset is a set of 28 28 gray-scale images each of which represents one of the following 10 fashion item classes... The CIFAR-10 dataset consists of 60,000 images in the 32 32 RGB format. |
| Dataset Splits | Yes | In this toy experiment, a training set, a validation set, a test set, and a dataset for estimating confidence were created from positive and negative samples drawn independently from two-dimensional Gaussian distributions. All the dataset except the validation set consisted of 1,000 positive samples and 1,000 negative samples. The validation set included only 1,000 positive samples. ... We then divided it into four stratified sub-datasets: a training set, a validation set, a test set, and a dataset for estimating confidence. |
| Hardware Specification | No | The paper does not provide specific hardware details (such as GPU/CPU models or memory) used for running its experiments. |
| Software Dependencies | No | The paper states 'We implemented the experiments on Python using Optuna [Akiba et al., 2019], PyTorch 2, and Scikit-learn [Pedregosa et al., 2011].', but it does not specify version numbers for PyTorch, Optuna, or Scikit-learn. |
| Experiment Setup | Yes | Adam [Kingma and Ba, 2015] with 5,000 epochs was used for optimization. We used a three-layer fully-connected neural network (784-100-100-1) with ReLU [Nair and Hinton, 2010] as an activation function and 10−7 weight decay. Optimization was executed by Adam [Kingma and Ba, 2015] for 200 epochs with minibatch size 100. ... Convolution (3 in/18 out-channels, kernel size 5). Max-pooling (kernel size 2, stride 2). Convolution (18 in/48 out-channels, kernel size 5). Max-pooling (kernel size 2, stride 2). Fully-connected (800 units) with ReLU. Fully-connected (400 units) with ReLU. Fully-connected (1 unit). |