Semantic-Guided Novel Category Discovery

Authors: Weishuai Wang, Ting Lei, Qingchao Chen, Yang Liu

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Through extensive experiments, we demonstrate the mutual benefits of the recognition and clustering tasks, which can be jointly optimized. Experimental results on multiple datasets confirm the effectiveness of our proposed method.
Researcher Affiliation Academia 1Wangxuan Institute of Computer Technology, Peking University 2National Institute of Health Data Science, Peking University
Pseudocode Yes Algorithm 1: Cluster-wise Pseudo-label Generation
Open Source Code Yes Our code is available at https://github.com/wang-weishuai/Semantic-guided-NCD.
Open Datasets Yes We evaluate our method on three benchmark NCD datasets: CIFAR10, CIFAR100 (Krizhevsky et al. 2009), and Image Net (Deng et al. 2009).
Dataset Splits Yes We follow the dataset splits of various settings in (Fini et al. 2021).
Hardware Specification No The paper does not specify any hardware details such as GPU models, CPU types, or memory used for the experiments. It only mentions 'The encoder fθ uses a Res Net18'.
Software Dependencies No The paper does not provide specific version numbers for any software dependencies or libraries. It only mentions 'Glove word vectors' and 'Res Net18'.
Experiment Setup Yes The encoder fθ uses a Res Net18 (He et al. 2016) pretrained on the labeled data set for the classification task. The weight α for the mutual information loss is set to 0.1. We choose K = 16 when we generate pseudo labels for unlabeled classes and use the Glove word vectors (Pennington, Socher, and Manning 2014) trained by Wikipedia2014 and Gigaword5 to supply semantic information.