Asymmetric Loss Functions for Learning with Noisy Labels

Authors: Xiong Zhou, Xianming Liu, Junjun Jiang, Xin Gao, Xiangyang Ji

ICML 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on benchmark datasets demonstrate that asymmetric loss functions can outperform state-of-the-art methods. The code is available at https://github.com/hitcszx/ALFs
Researcher Affiliation Academia 1Harbin Institute of Technology 2Peng Cheng Laboratory 3King Abdullah University of Science and Technology 4Tsinghua University.
Pseudocode No No pseudocode or algorithm block found in the paper.
Open Source Code Yes The code is available at https://github.com/hitcszx/ALFs
Open Datasets Yes In this section, we empirically investigate asymmetric loss functions on benchmark datasets, including MNIST (Lecun et al., 1998), CIFAR-10/-100 (Krizhevsky & Hinton, 2009) , and a real-world noisy dataset Web Vision (Li et al., 2017).
Dataset Splits No The top-1 validation accuracies under different loss functions on the clean Web Vision validation set are reported in Table 3.
Hardware Specification No No specific hardware details (e.g., GPU models, CPU types) are provided in the paper.
Software Dependencies No The paper mentions general software components like 'deep neural networks' and 'Res Net-50' but does not specify any software names with version numbers (e.g., PyTorch 1.9) needed to replicate the experiment.
Experiment Setup No The noise generation, networks, training details, hyper-parameter settings and more experimental results can be found in the supplementary material.