Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Socialized Coevolution: Advancing a Better World through Cross-Task Collaboration
Authors: Xinjie Yao, Yu Wang, Pengfei Zhu, Wanyu Lin, Ruipu Zhao, Zhoupeng Guo, Weihao Li, Qinghua Hu
ICML 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we compare DISC with state-of-the-art methods on the CIFAR100 and VOC07+12 datasets. We discuss task-driven knowledge transfer methods, and the ablation studies demonstrate the effectiveness of the DHC and DSC modules. Our model is implemented in PyTorch (Paszke et al., 2019) and deployed on an NVIDIA RTX 3090 GPU. |
| Researcher Affiliation | Academia | 1College of Intelligence and Computing, Tianjin University, Tianjin, China 2Engineering Research Center of City Intelligence and Digital Governance, Ministry of Education of the People s Republic of China, Tianjin, China 3Haihe Lab of ITAI, Tianjin, China 4Department of Computing, The Hong Kong Polytechnic University, Hong Kong, China 5School of Automation, Southeast University, Nanjing, China 6School of New Media and Communication, Tianjin University, Tianjin, China. Correspondence to: Pengfei Zhu <EMAIL>. |
| Pseudocode | Yes | For a clearer understanding of training and inference, we have described the detailed algorithms in Algorithms 1 and 2. |
| Open Source Code | Yes | Our code will be publicly available at https://github.com/yxjdarren/SC. |
| Open Datasets | Yes | Datasets: We evaluate the proposed method on CIFAR100 (Krizhevsky et al., 2009) and VOC07+12 (Everingham et al., 2010; 2015) datasets using a data-efficient setting. |
| Dataset Splits | Yes | Specifically, we employ 10% of the training set and the complete test set, with CIFAR100 used for evaluating classification and VOC07+12 for assessing detection... In the data-efficient setting, we utilize 10% of the training data: for VOC07+12 (10%), the training set contains 4,696 objects, and the testing set remains at 14,976 objects. For CIFAR100 (10%), the training set includes 5,000 images, with the testing set consisting of 10,000 images. |
| Hardware Specification | Yes | Our model is implemented in PyTorch (Paszke et al., 2019) and deployed on an NVIDIA RTX 3090 GPU. |
| Software Dependencies | No | Our model is implemented in PyTorch (Paszke et al., 2019) and deployed on an NVIDIA RTX 3090 GPU. This cites the PyTorch paper but does not provide a specific version number like "PyTorch 1.x". |
| Experiment Setup | Yes | For classification as the primary task, the model is trained for 1500 epochs using SGD with a batch size of 64. The threshold is set to 300, with a learning rate of 0.01, momentum of 0.9, and weight decay of 0.001. For detection as the primary task, the model is trained for 600 epochs using SGD with a batch size of 32. The threshold is also set to 300, with a learning rate of 0.03, momentum of 0.9, and weight decay of 0.001. |