Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Towards Reasonable Budget Allocation in Untargeted Graph Structure Attacks via Gradient Debias
Authors: Zihan Liu, Yun Luo, Lirong Wu, Zicheng Liu, Stan Z. Li
NeurIPS 2022 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate DEBIAS on three benchmark datasets against 7 baselines for untargeted graph structure attacks. The experimental results show that DEBIAS consistently outperforms baselines on both clean and poisoned graphs. |
| Researcher Affiliation | Academia | Department of Electrical and Computer Engineering, University of California, San Diego, La Jolla, CA, USA, 92093 |
| Pseudocode | Yes | Algorithm 1: Gradient Debiasing for Untargeted Graph Attacks |
| Open Source Code | No | The paper does not provide an explicit statement about releasing code or a link to a code repository for the methodology. |
| Open Datasets | Yes | Datasets We evaluate the effectiveness of DEBIAS on three widely-used benchmark datasets in graph machine learning: Cora, CiteSeer, and PubMed [19]. |
| Dataset Splits | Yes | For datasets Cora, CiteSeer, and PubMed, we use the commonly used 20/30/50 training/validation/testing split. |
| Hardware Specification | No | The paper does not explicitly describe the hardware used to run its experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details with version numbers (e.g., library or solver names with version numbers like Python 3.8, CPLEX 12.4) needed to replicate the experiment. |
| Experiment Setup | Yes | For all experiments, we use a two-layer GCN model [19] with hidden dimension 16 and a learning rate of 0.01. We use Adam optimizer [20] and train the GCN model for 200 epochs. To mitigate overfitting, we apply a dropout ratio of 0.5 for the GCN model. For datasets Cora, CiteSeer, and PubMed, we use the commonly used 20/30/50 training/validation/testing split. |