Robust Consensus Anchor Learning for Efficient Multi-view Subspace Clustering
Authors: Yalan Qin, Nan Pu, Guorui Feng, Nicu Sebe
ICML 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments performed on eight multiview datasets confirm the superiority of RCSC based on the effectiveness and efficiency. Extentive experiments on different multiview datasets validate the effectivenss and efficiency of RCSC, especially on the datasets with large scales. |
| Researcher Affiliation | Academia | 1 School of Communication and Information Engineering, Shanghai University, Shanghai 200444, China 2Department of Information Engineering and Computer Science, University of Trento, Trento 40128, Italy. |
| Pseudocode | Yes | Algorithm 1 Algorithm of RCSC |
| Open Source Code | No | The paper does not provide an explicit statement about the release of source code, nor does it include a link to a code repository or mention code in supplementary materials. |
| Open Datasets | Yes | For the experimental evaluation, we use eight real-world multi-view datasets, namely, ORL, Mfeat, Caltech101-20, Caltech101-all, SUNRGBD (Song et al., 2015), NUSWIDEOBJ (Chua et al., 2009), AWA and Youtube Face. |
| Dataset Splits | No | The paper does not provide specific details on how the datasets were split into training, validation, or test sets. It mentions using 'eight multi-view datasets' and repeating experiments but not the data partitioning methodology. |
| Hardware Specification | No | The paper does not provide specific details about the hardware specifications (e.g., GPU/CPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper does not provide specific details about ancillary software dependencies, such as library names with version numbers. |
| Experiment Setup | Yes | We first study how parameters β and λ influence the final clustering performance. These two parameters are adopted to negotiate the importances of partition term and Frobenious norm term. We illustrate the clustering performance of the proposed method with varying parameters λ and β in Fig. 1. It is observed that appropriate values for these two parameters are generally beneficial to the clustering results on different datasets. According to Fig. 1, we observe that relatively desired clustering results are achieved when β = 0.1 and λ = 0.1 on various datasets. Moreover, the results of the proposed method are generally stable over varying values within the range of parameters β and λ on different datasets, which shows that RCSC is generally robust to these two parameters. The anchor number of our method is tuned in the range of [2k, 3k, ..., 7k], where k denotes the total number of clusters in dataset. To reduce the randomness, we repeat each experiment for 20 times and report their mean values and variances in the experiment. |