Higher-Order Spectral Clustering of Directed Graphs
Authors: Steinar Laenen, He Sun
NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | The significance of our theoretical work is demonstrated by extensive experimental results on the UN Comtrade Dataset: the output clustering of our algorithm exhibits not only how the clusters (sets of countries) relate to each other with respect to their import and export records, but also how these clusters evolve over time, in accordance with known facts in international trade. The significance of our work is further demonstrated by experimental results on several synthetic and real-world datasets. |
| Researcher Affiliation | Collaboration | Steinar Laenen Oxford Research Group Five AI steinar.laenen@five.ai He Sun School of Informatics University of Edinburgh h.sun@ed.ac.uk |
| Pseudocode | No | The algorithm is described in a textual list of steps rather than a formally labeled pseudocode or algorithm block: "Our algorithm, which we call Simple Herm, only consists of a few lines of code and is described as follows: (1) compute the bottom eigenvector f1 Cn of the normalised Hermitian Laplacian matrix LG of G; (2) compute the embedding {F(v)}v V [G], where F(v) 1 dv f1(v) for any vertex v; (3) apply k-means on the embedded points {F(v)}v V [G]." |
| Open Source Code | No | The paper does not include any explicit statement about releasing source code or a link to a code repository for the described methodology. |
| Open Datasets | Yes | We compare our proposed algorithm against the previous state-of-the-art on the UN Comtrade Dataset [29]. This dataset consists of the import-export tradeflow data of 97 specific commodities across N = 246 countries and territories over the period 1964 2018. ... United Nations. UN comtrade free API. https://comtrade.un.org/data/. Accessed: 2020-06-03. |
| Dataset Splits | No | The paper does not explicitly provide details about training/validation/test splits for their datasets. It mentions running experiments over multiple years for the UN Comtrade dataset and averaging results over multiple runs for synthetic data, but no specific split ratios or methodology for typical model validation. |
| Hardware Specification | Yes | All our experiments are conducted with an ASUS Zen Book Pro UX501VW with an Intel(R) Core(TM) i7-6700HQ CPU @ 2.60GHz with 12GB of RAM. |
| Software Dependencies | No | The paper does not provide specific version numbers for any software dependencies. |
| Experiment Setup | Yes | Specifically, for given parameters k, n, p, q, η, a graph randomly chosen from the DSBM is constructed as follows: the overall graph consists of k clusters S0, . . . , Sk 1 of the same size... We set n = 1000, and k = 4. We set the value of p to be between 0.5 and 0.8, and the value of η to be between 0.5 and 0.7. As shown in Figure 1, our proposed Simple Herm clearly outperforms the Herm-RW and the DD-SYM algorithms. |