Multiple Clustering Views from Multiple Uncertain Experts
Authors: Yale Chang, Junxiang Chen, Michael H. Cho, Peter J. Castaldi, Edwin K. Silverman, Jennifer G. Dy
ICML 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results on synthetic datasets, benchmark datasets and a real-world disease subtyping problem show that our proposed approach outperforms competing baselines, including meta clustering, semisupervised clustering, semi-crowdsourced clustering and consensus clustering. |
| Researcher Affiliation | Academia | 1Northeastern University, Boston, MA 2Brigham and Women s Hospital, Harvard Medical School, Boston, MA. |
| Pseudocode | No | The paper describes its model and inference steps using mathematical formulations and descriptive text, but it does not include any explicitly labeled 'Pseudocode' or 'Algorithm' blocks. |
| Open Source Code | No | The paper does not provide any specific links to source code repositories or explicit statements about code availability for its proposed method. |
| Open Datasets | Yes | Web KB dataset (web, 1998) contains webpages collected from four universities. ... The Face dataset (Lichman, 2013) consists of 640 face images of people taken with varying poses... |
| Dataset Splits | Yes | We randomly split the dataset into half training set and half testing set. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., CPU, GPU models, memory) used to run the experiments. |
| Software Dependencies | No | The paper mentions tools like 'Autograd: Reverse-mode differentiation of native python' (Maclaurin et al., 2015) and the use of 'a limited-memory projected quasi-Newton algorithm (PQN)' (Schmidt et al., 2009), but it does not provide specific version numbers for the software dependencies of its implementation. |
| Experiment Setup | Yes | We provide the parameter setting details for all methods in the supplementary materials due to space constraint. ... We set the number of random initializations to be 50 in all experiments and the results are stable across different runs. ... We simulate noisy constraints provided by multiple expert from two latent views, Y1 and Y2 as follows: 1) the first view consists of experts 1-5, who provide constraints based on clustering solution Y1 and have accuracy parameters α1:5 = β1:5 = (0.95, 0.9, 0.85, 0.8, 0.75); ... the number of constraints, ncon, is varied from 200 to a large number... |