DIMES: A Differentiable Meta Solver for Combinatorial Optimization Problems
Authors: Ruizhong Qiu, Zhiqing Sun, Yiming Yang
NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments show that DIMES outperforms recent DRL-based methods on large benchmark datasets for Traveling Salesman Problems and Maximal Independent Set problems. 4 Experiments 4.1 Experiments for Traveling Salesman Problem 4.1.1 Experimental Settings |
| Researcher Affiliation | Academia | Ruizhong Qiu Department of Computer Science University of Illinois Urbana Champaign rq5@illinois.edu Zhiqing Sun , Yiming Yang Language Technologies Institute Carnegie Mellon University {zhiqings,yiming}@cs.cmu.edu |
| Pseudocode | Yes | Algorithm 1 MAML in DIMES |
| Open Source Code | Yes | Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [Yes] See Appendix C. |
| Open Datasets | Yes | The training instances are generated on the fly. We closely follow the data generation procedure of previous works, e.g., [42]. For testing, we use the test instances generated by Fu et al. [19]. We mainly focus on two types of graphs that recent work [48, 1, 8] shows struggles against, i.e., Erd os-Rényi (ER) graphs [16] and SATLIB [25]. |
| Dataset Splits | No | The paper mentions 'training instances' and 'test instances' for TSP and '4,096 training and 5,00 test ER graphs' and 'train on 39,500 and test on 500' for SATLIB, but it does not explicitly specify a validation dataset split or percentages used for validation. |
| Hardware Specification | Yes | Training and Hardware Due to the space limit, please refer to the appendix. Did you include the total amount of compute and the type of resources used (e.g., type of GPUs, internal cluster, or cloud provider)? [Yes] See Appendix C. |
| Software Dependencies | No | The paper states 'Training and Hardware Due to the space limit, please refer to the appendix.' and 'Did you specify all the training details (e.g., data splits, hyperparameters, how they were chosen)? [Yes] See Appendix C.'. While this implies training details are provided, it does not explicitly state that specific software dependencies with version numbers are included. |
| Experiment Setup | Yes | Training and Hardware Due to the space limit, please refer to the appendix. Did you specify all the training details (e.g., data splits, hyperparameters, how they were chosen)? [Yes] See Appendix C. |