InfoOT: Information Maximizing Optimal Transport
Authors: Ching-Yao Chuang, Stefanie Jegelka, David Alvarez-Melis
ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Empirically, Info OT improves the quality of alignments across benchmarks in domain adaptation, cross-domain retrieval, and single-cell alignment. The code is available at https: //github.com/chingyaoc/Info OT. and 6. Experiments We now evaluate Info OT with experiments in point cloud matching, domain adaptation, cross-domain retrieval, and single-cell alignment. |
| Researcher Affiliation | Collaboration | 1MIT CSAIL, Cambridge, MA, USA 2Microsoft Research, Cambridge, MA, USA. |
| Pseudocode | No | The paper describes algorithms and derivations but does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | The code is available at https: //github.com/chingyaoc/Info OT. |
| Open Datasets | Yes | We apply the fused version of Info OT to two domain adaptation benchmarks: MNIST-USPS and the Office Caltech dataset (Gong et al., 2012). |
| Dataset Splits | Yes | Following (Flamary et al., 2016), we present the results over 10 independent trials. In each trial of Office-Caltech, the target data is divided into 90%/10% train-test split, where OT and 1-NN classifiers are only computed on the training set. For MNIST-USPS, only 2000 samples from the source and target training set are used, while the original test sets are used. ... The bandwidth for each benchmark is selected from {0.2, 0.3, ..., 0.8} with the circular validation procedure (Bruzzone & Marconcini, 2009; Perrot et al., 2016; Zhong et al., 2010) on M U and A D, which is 0.4 and 0.5, respectively. |
| Hardware Specification | No | The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments. |
| Software Dependencies | No | All the optimal transport approaches are implemented or adopted from the POT library (Flamary et al., 2021). The paper does not specify version numbers for this or any other software used. |
| Experiment Setup | Yes | The strength of the entropy regularizer ϵ is set to 1 for every entropic regularized OT, and the λ of FInfo OT is set to 100 for all the experiments. The bandwidth for each benchmark is selected from {0.2, 0.3, ..., 0.8} with the circular validation procedure (Bruzzone & Marconcini, 2009; Perrot et al., 2016; Zhong et al., 2010) on M U and A D, which is 0.4 and 0.5, respectively. |