Data Compression for Learning MRF Parameters
Authors: Khaled S. Refaat, Adnan Darwiche
IJCAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | The experimental results are given in Section 5. |
| Researcher Affiliation | Academia | Khaled S. Refaat and Adnan Darwiche Computer Science Department University of California, Los Angeles {krefaat,darwiche}@cs.ucla.edu |
| Pseudocode | No | The paper describes the data decomposition process in prose but does not include any structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not include an unambiguous statement that the authors are releasing the code for the described methodology, nor does it provide a link to a source-code repository. |
| Open Datasets | No | The paper states 'we simulate a dataset' and refers to network structures, but does not provide concrete access information (link, DOI, specific repository, or formal citation for public availability) for the specific datasets used in the experiments. While it references the UCI repository, it does so as an example of dataset characteristics, not as a direct source for their experimental data. |
| Dataset Splits | No | The paper does not provide specific dataset split information (exact percentages, sample counts, or citations to predefined splits) for training, validation, or testing. It mentions 'simulated a dataset' and 'hiding 20% of the variables' but no clear partitioning. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory amounts, or cloud instance types) used for running its experiments. It only vaguely mentions 'low-cost computers'. |
| Software Dependencies | No | The paper mentions using 'gradient descent and EM' and comparing against 'Fast Inf', but it does not provide specific version numbers for any programming languages, libraries, or frameworks used in their implementation. |
| Experiment Setup | Yes | In particular, using a fixed network structure, we simulate a dataset, then make the data incomplete by randomly selecting a certain percentage of variables to have missing values. After that, we learn the parameters from the data using the gradient method with and without data decomposition, to obtain a local optimum. For 11 different networks, 3 and with hiding 20% of the variables, Table 1 shows... Figure 4 shows the speed-up obtained by our Gradient and EM methods, that use data decomposition (allowed 100 iterations), over Fast Inf EM (with the gradient option allowed only 2 iterations)... |