Predictive Flows for Faster Ford-Fulkerson
Authors: Sami Davies, Benjamin Moseley, Sergei Vassilvitskii, Yuyan Wang
ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We then consider image segmentation, a common use-case of flows in computer vision, and complement our theoretical analysis with strong empirical results. |
| Researcher Affiliation | Collaboration | 1Department of Computer Science, Northwestern University, Evanston IL, USA 2Tepper School of Business, Carnegie Mellon University, Pittsburgh PA, USA 3Google Research, NYC NY, USA. |
| Pseudocode | Yes | Algorithm 1 Warm-starting Ford-Fulkerson with bf |
| Open Source Code | Yes | The omitted data tables and other experiment results can be found in the uploaded program directory (see the README.md file for instructions). |
| Open Datasets | Yes | We use four different image groups from the Pattern Recognition and Image Processing dataset from the University of Freiburg4, named BIRDHOUSE, HEAD, SHOE and DOG respectively. |
| Dataset Splits | No | The paper describes data cropping and resizing, but it does not explicitly provide details about training, validation, or test dataset splits (e.g., percentages or sample counts for each partition). |
| Hardware Specification | Yes | All experiments are run on a device with Intel(R) Core(TM) i7-7600U CPU @ 2.80GHz, with 24G memory. |
| Software Dependencies | No | The paper mentions that 'Many of the image process tools and functions are based on the Image Segmentation Github project (Jiang, 2017)', but it does not specify exact version numbers for any software dependencies. |
| Experiment Setup | Yes | Numerically, the σ in the definition of β , is 50, and C is 100. To make the capacities integral, all βp,q s are rounded down to the nearest integer. Notice that βp,q C by definition. We let M = C|V |2 to make the term sufficiently large. |