Scaling Up Semi-supervised Learning with Unconstrained Unlabelled Data
Authors: Shuvendu Roy, Ali Etemad
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We perform extensive experiments on 4 commonly used datasets and demonstrate superior performance over existing semi-supervised methods with a performance boost of 4.79%. Extensive ablation and sensitivity studies show the effectiveness and impact of each of the proposed components of our method. |
| Researcher Affiliation | Academia | Shuvendu Roy, Ali Etemad Queen s University, Canada {shuvendu.roy, ali.etemad}@queensu.ca |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks. It provides a diagrammatic overview of the method in Figure 2 and describes the mathematical formulations. |
| Open Source Code | Yes | The code for our work is publicly available. To facilitate reproducibility, we release the code at github.com/Shuvendu Roy/Un Mix Match. |
| Open Datasets | Yes | For our main experiments, we follow the standard semi-supervised evaluation protocol from prior works (Sohn et al. 2020), and present the results for four datasets: CIFAR-10 (Krizhevsky, Hinton et al. 2009), CIFAR-100 (Krizhevsky, Hinton et al. 2009), SVHN (Netzer et al. 2011), and STL-10 (Coates, Ng, and Lee 2011). We use Image Net-1K (Deng et al. 2009), and Image Net-100 (a subset of Image Net-1K) as the unconstrained unlabeled datasets... |
| Dataset Splits | No | The paper does not explicitly state a specific validation dataset split. It mentions 'different numbers of labelled samples' and 'averaged over three runs', and refers to 'standard semi-supervised evaluation protocol from prior works (Sohn et al. 2020)', but a distinct validation set split is not detailed. |
| Hardware Specification | No | The paper states: 'We are also thankful to Sci Net HPC Consortium for helping with the computation resources,' but does not provide specific hardware details such as GPU models, CPU models, or memory amounts used for the experiments. |
| Software Dependencies | No | The paper states: 'The code is implemented with Pytorch and built using Torch SSL (Zhang et al. 2021),' but does not provide specific version numbers for Pytorch or Torch SSL, nor other ancillary software with versions. |
| Experiment Setup | Yes | We train the method for 220 iterations with a batch size of 64, a learning rate of 0.03, and an SGD optimizer with a momentum of 0.9 and weight-decay of 0.0005. For the encoder, following existing literature such as (Sohn et al. 2020; Zhang et al. 2021), we use Wide Res Net-28-2 (Zagoruyko and Komodakis 2016) for CIFAR-10 and SVHN, WRN-28-8 for CIFAR-100, and WRN-37-2 for STL-10. |