Label-Agnostic Forgetting: A Supervision-Free Unlearning in Deep Models
Authors: Shaofei Shen, Chenhao Zhang, Yawen Zhao, Alina Bialkowski, Weitong Tony Chen, Miao Xu
ICLR 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we conduct experiments to answer three research questions to evaluate LAF: RQ1: How do the proposed LAF perform on the data removal, class removal, and noisy label removal tasks, as compared with the state-of-the-art unlearning methods? RQ2: Is the representation space after the implementation of LAF consistent with the space of retrained models? RQ3: How do LUE and LRA affect the performance of LAF? [...] Table 1, 2, and 3 showcase the experiment results of the three distinct tasks: data removal, class removal, and noisy label removal. |
| Researcher Affiliation | Academia | 1University of Queensland 2University of Adelaide {shaofei.shen,chenhao.zhang,yawen.zhao,alina.bialkowski}@uq.edu.au t.chen@adelaide.edu.au, miao.xu@uq.edu.au |
| Pseudocode | Yes | We provide the pseudo-code of LAF in Algorithm 1. |
| Open Source Code | Yes | 1https://github.com/Shaofei Shen768/LAF |
| Open Datasets | Yes | Datasets and Models. To validate the effectiveness of LAF, we conduct experiments on four datasets: DIGITS (MNIST) (Le Cun, 1998), FASHION (Fashion MNIST)(Xiao et al., 2017), CIFAR10 (Krizhevsky et al., 2009) and SVHN (Netzer et al., 2011). |
| Dataset Splits | No | The paper describes various data categories like training data (D), remaining data (Dr), and forgetting data (Df) and evaluation metrics on 'Trainr' and 'Test', but it does not provide explicit, reproducible percentages or counts for distinct training, validation, and test dataset splits required for reproduction. |
| Hardware Specification | Yes | All the experiments are conducted on one server with NVIDIA RTX A6000 GPU (48GB GDDR6 Memory) and 12th Gen Intel(R) Core(TM) i9-12900K (16 cores and 128GB Memory) and two servers with NVIDIA RTX A5000 GPUs (24GB GDDR6 Memory) and 12th Gen Intel Core i7-12700K CPUs (12 cores and 128GB Memory). |
| Software Dependencies | Yes | The code of LAF was implemented in Python 3.9.16 and Cuda 11.6.1. The main Python packages versions are the following: Numpy 1.23.5; Pandas 2.0.1; Pytorch 1.13.1; Torchvision 0.14.1. |
| Experiment Setup | Yes | In all experiments, we configure the batch size to 32. During the training of VAEs, we assign the latent dimensions as 8 for the DIGITS and FASHION datasets and 16 for the CIFAR10 and SVHN datasets. The learning rate for VAE training is established at 1e-3, with the number of training epochs set to 10. For representation alignment, we assign the value of τ as 2, 20, and 20 for data removal, class removal, and noisy label removal tasks, respectively for CNN. We assign the value of τ as 20, 20, and 5 for Res Net. Subsequently, in the supervised repairing stage, we designate the repairing epoch as 1, applying a learning rate of 1e-3 for all tasks on the DIGITS and FASHION datasets, and 5e-5 on the CIFAR10 and SVHN datasets. |