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..
Global graph kernels using geometric embeddings
Authors: Fredrik Johansson, Vinay Jethava, Devdatt Dubhashi, Chiranjib Bhattacharyya
ICML 2014 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate the Lov asz ϑ and SVM-ϑ kernels by performing graph classification of synthetic and benchmark datasets. We report the classification accuracy using 10fold cross-validation with a C-Support Vector Machine classifier, LIBSVM (Chang & Lin, 2011). |
| Researcher Affiliation | Academia | Fredrik D. Johansson EMAIL Chalmers University of Technology, SE-412 96 Gothenburg, Sweden Vinay Jethava EMAIL Chalmers University of Technology, SE-412 96 Gothenburg, Sweden Devdatt Dubhashi EMAIL Chalmers University of Technology, SE-412 96 Gothenburg, Sweden Chiranjib Bhattacharyya EMAIL Indian Institute of Science, Bangalore 560012 Karnataka, India |
| Pseudocode | No | The paper describes methods but does not include any formally structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide an explicit statement or link to its own open-source code for the described methodology. |
| Open Datasets | Yes | PTC (Predictive Toxicology Challenge) is a set of 417 chemical compound graphs labeled according to their carcinogenic effects on rats and mice (Helma et al., 2001). MUTAG (Debnath et al., 1991) is a dataset of 188 graphs representing mutagenetic compounds... ENZYME is a collection of 600 graphs representing tertiary protein structures collected by (Borgwardt et al., 2005)... NCI1 is a set of 4110 graphs representing a subset of chemical compounds screened for activity against non-small cell lung cancer cell lines (Wale et al., 2008). |
| Dataset Splits | Yes | We report the classification accuracy using 10fold cross-validation with a C-Support Vector Machine classifier, LIBSVM (Chang & Lin, 2011). |
| Hardware Specification | Yes | We report the CPU runtimes for computing each kernel on the benchmark experiments in Table 2, as measured in Matlab R2012a on a 3.4GHz Intel Core i7 with 4 cores and 32GB RAM. |
| Software Dependencies | Yes | as measured in Matlab R2012a on a 3.4GHz Intel Core i7 with 4 cores and 32GB RAM. ... LIBSVM (Chang & Lin, 2011). |
| Experiment Setup | Yes | The SVM parameter C was optimized for each kernel and fold and the best was used for the final accuracy. ... Both kernels were used with either the linear kernel k(x,y) = xy or the radial basis function kernel k(x,y) = e x y 2 2/(2σ2) with σ from the set [0.01,10]. |