Learning Mixtures of Plackett-Luce Models
Authors: Zhibing Zhao, Peter Piech, Lirong Xia
ICML 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our experiments show that our GMM algorithm is significantly faster than the EMM algorithm by Gormley & Murphy (2008), while achieving competitive statistical efficiency. |
| Researcher Affiliation | Academia | Zhibing Zhao ZHAOZ6@RPI.EDU Rensselaer Polytechnic Institute, 110 8th St., Troy, NY 12180 USA Peter Piech PIECHP@RPI.EDU Rensselaer Polytechnic Institute, 110 8th St., Troy, NY 12180 USA Lirong Xia XIAL@CS.RPI.EDU Rensselaer Polytechnic Institute, 110 8th St., Troy, NY 12180 USA |
| Pseudocode | Yes | Algorithm 1 GMM for 2-PL Input: Preference profile P with n full rankings. Compute the frequency of each of the 20 moments Compute the output according to (4) |
| Open Source Code | No | The paper does not provide any statement or link indicating that the source code for the methodology is openly available. |
| Open Datasets | No | The paper uses 'synthetic data' that is generated according to a described procedure rather than a publicly available dataset. No access information for a dataset is provided. |
| Dataset Splits | No | The paper does not explicitly provide details about train/validation/test dataset splits. It mentions generating synthetic datasets for evaluation, but not specific partitioning for training, validation, and testing within those datasets. |
| Hardware Specification | Yes | All experiments are run on an Ubuntu Linux server with Intel Xeon E5 v3 CPUs each clocked at 3.50 GHz. |
| Software Dependencies | Yes | The GMM algorithm is implemented in Python 3.4 and termination criteria for the optimization are convergence of the solution and the objective function values to within 10^-10 and 10^-6 respectively. The optimization of (4) uses the fmincon function through the MATLAB Engine for Python. The EMM algorithm is also implemented in Python 3.4 and the E and M steps are repeated together for a fixed number of iterations without convergence criteria. |
| Experiment Setup | Yes | The GMM algorithm is implemented in Python 3.4 and termination criteria for the optimization are convergence of the solution and the objective function values to within 10^-10 and 10^-6 respectively. The EMM algorithm is also implemented in Python 3.4 and the E and M steps are repeated together for a fixed number of iterations without convergence criteria. We have tested all configurations of EMM with 10 and 20 overall MM iterations, respectively. We found that the optimal configurations are EMM-10-1 and EMM-20-1 (shown in Figure 1, results for other configurations are omitted), where EMM-20-1 means 20 iterations of E step, each of which uses 1 MM iteration. |