A Graph Theoretic Additive Approximation of Optimal Transport

Authors: Nathaniel Lahn, Deepika Mulchandani, Sharath Raghvendra

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We also provide empirical results that suggest our algorithm is competitive with respect to a sequential implementation of the Sinkhorn algorithm in execution time. Moreover, our algorithm quickly computes a solution for very small values of δ whereas Sinkhorn algorithm slows down due to numerical instability.
Researcher Affiliation Academia Nathaniel Lahn Department of Computer Science Virginia Tech Blacksburg, VA 24061 lahnn@vt.edu Deepika Mulchandani Virginia Tech Blacksburg, VA 24061 deepikak@vt.edu Sharath Raghvendra Virginia Tech Blacksburg, VA 24061 sharathr@vt.edu
Pseudocode No The algorithm's steps are described in detail within the text (e.g., Section 2.1 'The algorithm', 'First step (Hungarian Search)'), but there is no formally labeled pseudocode block or algorithm listing.
Open Source Code Yes Our implementation is available at https://github.com/nathaniellahn/Combinatorial Optimal Transport.
Open Datasets Yes All the tests are conducted on real-world data generated by randomly selecting pairs of images from the MNIST data set of handwritten digits.
Dataset Splits No The paper states it uses 'randomly selecting pairs of images from the MNIST data set' for its tests, but does not specify explicit train, validation, or test dataset splits or a partitioning methodology for the data.
Hardware Specification Yes All tests are executed on computer with a 2.40 GHz Intel Dual Core i5 processor and 8GB of RAM, using a single computation thread.
Software Dependencies No The paper states the implementation is 'written in Java' and the 'testing code is written in MATLAB', but it does not specify version numbers for these languages or any other software dependencies like libraries or solvers.
Experiment Setup Yes We implement a worst-case asymptotically optimal algorithm as presented in this paper and fix the value of ε as 0.5.