Diversity-enhancing Generative Network for Few-shot Hypothesis Adaptation
Authors: Ruijiang Dong, Feng Liu, Haoang Chi, Tongliang Liu, Mingming Gong, Gang Niu, Masashi Sugiyama, Bo Han
ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results show that the DEG-Net outperforms existing FHA baselines and further verifies that generating diverse data plays a vital role in addressing the FHA problem. We verify the effectiveness of DEG-Net on 8 FHA benchmark tasks (Chi et al., 2021a), including the digit datasets and object datasets. Experimental results show that DEG-Net outperforms existing FHA methods and achieves stateof-the-art performance. |
| Researcher Affiliation | Academia | 1Department of Computer Science, Hong Kong Baptist University 2School of Mathematics and Statistics, The University of Melbourne 3Center for Advanced Intelligence Project, RIKEN 4Tianjin Artificial Intelligence Innovation Center 5Mohamed bin Zayed University of Artificial Intelligence 6Sydney AI Centre, The University of Sydney 7Graduate School of Frontier Sciences, The University of Tokyo. |
| Pseudocode | Yes | Algorithm 1 Diversity-enhancing Generative Network (DEG-Net) |
| Open Source Code | No | The paper does not provide any statement or link indicating the availability of open-source code for the described methodology. |
| Open Datasets | Yes | We compare DEG-Net with previous FHA methods on digits datasets (i.e. MNIST (M), USPS (U), and SVHN (S)) and objects datasets (i.e. CIFAR-10 (CF) , STL-10 (SL) and Vis DA-C), following (Chi et al., 2021a). MNIST (M) (Le Cun et al., 1998) is the handwritten digits dataset... SVHN (S) (Netzer et al., 2011) is the real-world image digits dataset... USPS (U) (Hull, 1994) data are 16 16 grayscale pixels. The CIFAR-10 (Krizhevsky et al., 2009) dataset contains 60, 000 32 32 color images... STL-10 (Coates et al., 2011) dataset is inspired by the CIFAR-10 dataset... Vis DA-C (Peng et al., 2017), which is a challenging large-scale datasets... |
| Dataset Splits | Yes | We conduct 6 tasks of the adaptation among the 3 digital datasets and choose the number of target data from 1 to 7 per class. Following Chi et al. (2021a) and Motiian et al. (2017), We conduct 6 tasks of the adaptation among the 3 digital datasets and choose the number of target data from 1 to 7 per class. |
| Hardware Specification | Yes | We implement all methods by Py Torch 1.7.1 and Python 3.7.6, and conduct all the experiments on NVIDIA RTX 2080Ti GPUs. |
| Software Dependencies | Yes | We implement all methods by Py Torch 1.7.1 and Python 3.7.6, and conduct all the experiments on NVIDIA RTX 2080Ti GPUs. |
| Experiment Setup | Yes | We pretrain the conditional generator for 300 epochs and pretrain the group discriminator for 100 epochs. The training step of the classifier (i.e. the adaptation module) is set to 50. As for the generator and the group discriminator, the learning rate of adam optimizer is set to 1 10 4. As for the classifier, the learning rate of adam optimizer is set to 1 10 3. The tradeoff parameter λ in Eq. (8) is set to 0.9 and the tradeoff parameter β in Eq. (11) is set to 0.1. Following Long et al. (2018) the radeoff parameter γ in Eq. (13) is set to 2 1+exp( 10 q) 1. |