Learning on graphs using Orthonormal Representation is Statistically Consistent

Authors: Rakesh Shivanna, Chiranjib Bhattacharyya

NeurIPS 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct two sets of experiments. ... Table 1 summarizes the results. Each entry is accuracy in % w.r.t. 0-1 loss, and the results were averaged over 100 iterations.
Researcher Affiliation Academia Rakesh S Department of Electrical Engineering Indian Institute of Science Bangalore, 560012, INDIA rakeshsmysore@gmail.com Chiranjib Bhattacharyya Department of CSA Indian Institute of Science Bangalore, 560012, INDIA chiru@csa.iisc.ernet.in
Pseudocode Yes Algorithm 1 Input: U, y S and C > 0. Get α , y S by solving ΛC(K, y S) (2) for ℓhinge and K = U U. Return: ˆy = U h S, where h S = UYα ; Y Dn , Y = yi, if i S, otherwise y i .
Open Source Code No No explicit statement providing concrete access to source code for the methodology. Footnote 9 mentions 'Relevant resources at: mllab.csa.iisc.ernet.in/rakeshs/nips14' which is too general to confirm code availability.
Open Datasets Yes We use two datasets similarity matrices from [11] and RBF kernel10 as similarity matrices for the UCI datasets [8]. ... We investigate the recently launched Google dataset [17], which contains multiple views of video game You Tube videos
Dataset Splits No The paper states '10% of labelled nodes observable' and 'We considered 20% of the data to be labelled' for the input, and 'We choose the parameters λ, λ1 and λ2 by cross validation.' but does not provide specific training/validation/test dataset splits (e.g., exact percentages or sample counts for each partition).
Hardware Specification No The paper does not provide any specific hardware details (e.g., GPU/CPU models, memory, or cloud instance types) used for running its experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., Python, PyTorch, TensorFlow, or specific library versions) needed to replicate the experiment.
Experiment Setup No The paper states that parameters (λ, λ1, λ2) were chosen by cross-validation, but it does not provide the specific hyperparameter values or other concrete experimental setup details such as learning rates, batch sizes, or optimizer settings.