NeuPSL: Neural Probabilistic Soft Logic

Authors: Connor Pryor, Charles Dickens, Eriq Augustine, Alon Albalak, William Yang Wang, Lise Getoor

IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Through an extensive empirical evaluation, we demonstrate the benefits of using Ne Sy methods, achieving upwards of 30% improvement over independent neural network models. We perform extensive evaluations over two image classification tasks and two citation network datasets.
Researcher Affiliation Academia Connor Pryor1 , Charles Dickens1 , Eriq Augustine 1 , Alon Albalak 2 , William Yang Wang 2 and Lise Getoor 1 1 UC Santa Cruz 2 UC Santa Barbara
Pseudocode No The paper does not contain any structured pseudocode blocks or algorithms explicitly labeled as such.
Open Source Code Yes Code and Data: https://github.com/linqs/neupsl-ijcai23
Open Datasets Yes The first set of experiments are conducted on a variation of MNIST Addition, a widely used Ne Sy evaluation task [Manhaeve et al., 2018]. Inspired by the Visual Sudoku problem proposed by Wang et al. (2019), Augustine et al. (2022) introduced a novel Ne Sy task, Visual-Sudoku-Classification. In our final experiment, we evaluate the performance of Neu PSL on two widely studied citation network node classification datasets: Citeseer and Cora [Sen et al., 2008].
Dataset Splits Yes We averaged the results over ten randomly sampled splits using 5% of the nodes for training, 5% of the nodes for validation, and 1000 nodes for testing.
Hardware Specification No The paper states that "Implementation details, hyperparameters, network architectures, hardware, and Neu PSL models, are described in the Appendix." and provides a link to the appendix, but the specific hardware details are not present within the provided text.
Software Dependencies No The paper mentions using "the open-source PSL software package" and "TensorFlow" but does not provide specific version numbers for these software dependencies, which are necessary for reproducibility.
Experiment Setup No The paper states that "Implementation details, hyperparameters, network architectures, hardware, and Neu PSL models, are described in the Appendix." and provides a link to the appendix, but the specific experimental setup details (e.g., hyperparameters) are not present within the provided text.