Learning General Latent-Variable Graphical Models with Predictive Belief Propagation

Authors: Borui Wang, Geoffrey Gordon6118-6126

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate its performance on both synthetic and real datasets, and showed that it learns different types of latent graphical models efficiently and achieves superior inference performance.
Researcher Affiliation Academia Borui Wang Computer Science Department Stanford University wbr@cs.stanford.edu Geoffrey Gordon Machine Learning Department Carnegie Mellon University ggordon@cs.cmu.edu
Pseudocode Yes See Appendix A.3 for a pseudocode summary of our algorithm and see Appendix A.5 for the proof of consistency of our algorithm.
Open Source Code No The paper does not provide a direct link to its source code or explicitly state that the code for their method is open-source or publicly available.
Open Datasets Yes Pen-Based Recognition of Handwritten Digits dataset in the UCI machine learning repository (Asuncion and Newman 2007).
Dataset Splits Yes Here we use 7000 samples as our training set, 494 samples as our validation set, and the other 3498 samples as our testing set.
Hardware Specification No The paper does not specify any particular hardware components such as GPU or CPU models used for the experiments.
Software Dependencies No The paper mentions using 'ridge regression' but does not specify the version numbers of any software libraries or dependencies used (e.g., Python, PyTorch, scikit-learn versions).
Experiment Setup Yes In our experiment, we use Gaussian radial basis function kernel embeddings with bandwidth parameter σ = 10 as our feature vectors, and use ridge regression (Friedman, Hastie, and Tibshirani 2009) with regularization parameter λ = 0.1 for S1A and S1B.