CuCo: Graph Representation with Curriculum Contrastive Learning

Authors: Guanyi Chu, Xiao Wang, Chuan Shi, Xunqiang Jiang

IJCAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on fifteen graph classification real-world datasets, as well as the parameter analysis, well demonstrate that our proposed Cu Co yields truly encouraging results in terms of performance on classification and convergence.
Researcher Affiliation Academia Guanyi Chu , Xiao Wang , Chuan Shi and Xunqiang Jiang Beijing University of Posts and Telecommunications {cgy463, xiaowang, shichuan, skd621}@bupt.edu.cn
Pseudocode Yes Algorithm 1 Training procedure of Cu Co
Open Source Code No The paper does not provide any explicit statement about open-sourcing its code, nor does it include a link to a code repository.
Open Datasets Yes We evaluate model performance on seven classical graph classification benchmarks shown in Table 1... we evaluate model performance on eight Open Graph Benchmark (OGB) [Weihua et al., 2020b] molecule property prediction datasets.
Dataset Splits Yes We use 10-fold cross validation accuracy to report the classification performance.
Hardware Specification No The paper does not provide specific hardware details (e.g., CPU/GPU models, memory specifications) used for running the experiments.
Software Dependencies No The paper mentions employing 'graph neural networks (GNNs)' and specifically 'a three-layer Graph Isomorphism Network (GIN)', but it does not specify software dependencies with version numbers (e.g., PyTorch 1.x, TensorFlow 2.x, Python 3.x).
Experiment Setup Yes For our proposed model, we adopt a three-layer Graph Isomorphism Network (GIN) with 32-dimensional hidden units and a sum pooling readout function for performance comparisons. We use 10-fold cross validation accuracy to report the classification performance. Experiments are repeated 5 times.