Graph Structured Prediction Energy Networks

Authors: Colin Graber, Alexander Schwing

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We apply the proposed method to tasks from the natural language processing and computer vision domain and demonstrate its general utility. To demonstrate the utility of our model and to compare inference and learning settings, we report results on the tasks of optical character recognition (OCR), image tagging, multilabel classification, and named entity recognition (NER).
Researcher Affiliation Academia Colin Graber Alexander Schwing cgraber2@illinois.edu aschwing@illinois.edu Department of Computer Science University of Illinois at Urbana-Champaign Champaign, IL
Pseudocode Yes Algorithm 1 Frank-Wolfe Inference for GSPEN
Open Source Code Yes Our implementation is available at https://github.com/cgraber/GSPEN.
Open Datasets Yes For the OCR experiments... from the Chars74k dataset [20], evaluate on the MIRFLICKR25k dataset [21], We use the Bibtex and Bookmarks multilabel datasets [23], Co NLL 2003 shared task [25]
Dataset Splits Yes The training, validation, and test set sizes for each dataset are 10,000, 2,000, and 2,000, respectively.
Hardware Specification No The paper mentions 'NVIDIA for providing GPUs' and 'Cisco for access to the Arcetri cluster' but does not specify exact GPU models, CPU models, or other detailed hardware specifications.
Software Dependencies No The paper does not provide specific software dependency versions (e.g., library or framework versions like PyTorch 1.9, TensorFlow 2.x).
Experiment Setup Yes Additional experimental details, including hyper-parameter settings, are provided in Appendix A.2.