Mixture Proportion Estimation via Kernel Embeddings of Distributions
Authors: Harish Ramaswamy, Clayton Scott, Ambuj Tewari
ICML 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We run our algorithm on several standard classification datasets, and demonstrate that it performs comparably to or better than other algorithms on most datasets. |
| Researcher Affiliation | Collaboration | Harish G. Ramaswamy HGRAMASW@IN.IBM.COM IBM Research, Bangalore, India Indian Institute of Science, Bangalore, India Clayton Scott CLAYSCOT@UMICH.EDU EECS and Statistics, University of Michigan, Ann Arbor, USA Ambuj Tewari TEWARIA@UMICH.EDU Statistics and EECS, University of Michigan, Ann Arbor, USA |
| Pseudocode | Yes | Algorithm 1 Kernel mean based gradient thresholder |
| Open Source Code | Yes | The code for our algorithms KM1 and KM2 are at http: //web.eecs.umich.edu/ cscott/code.html#kmpe. |
| Open Datasets | Yes | We ran our algorithm with 6 standard binary classification datasets taken from the UCI machine learning repository, the details of which are given below in Table 1. |
| Dataset Splits | No | The paper does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) needed to reproduce the data partitioning for training, validation, or testing. |
| Hardware Specification | No | The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments. |
| Software Dependencies | No | The paper mentions 'general purpose convex programming solver CVXOPT' but does not provide specific version numbers for any software dependencies. |
| Experiment Setup | Yes | Algorithm 1 lists constants like ' = 0.04, λUB = 10'. The paper also states: 'Our candidate kernels were five Gaussian RBF kernels, with the kernel width taking values uniformly in the log space between a tenth of the median pairwise distance and ten times the median distance, and among these kernels the kernel for which kφ( b F) φ( b H)k is highest is chosen.' |