Towards Enabling Learnware to Handle Unseen Jobs

Authors: Yu-Jie Zhang, Yu-Hu Yan, Peng Zhao, Zhi-Hua Zhou10964-10972

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Empirical studies also validate the efficacy of our approach. We also validate the effectiveness of our proposal by extensive experiments. This section examines the efficacy of our method, where we compare our method with contenders in various scenarios.
Researcher Affiliation Academia Yu-Jie Zhang, Yu-Hu Yan, Peng Zhao, Zhi-Hua Zhou National Key Laboratory for Novel Software Technology, Nanjing University, Nanjing 210023, China {zhangyj, yanyh, zhaop, zhouzh}@lamda.nju.edu.cn
Pseudocode Yes Detailed implementations of the MPE estimator (Algorithm 1) and selector training (Algorithm 2) are provided in Appendix A.
Open Source Code No The paper does not provide an explicit statement or link for the open-sourcing of the code for the described methodology.
Open Datasets Yes The evaluation is conducted on three widely used benchmark datasets: CIFAR-100 (Krizhevsky 2009), Newsgroup20 (Joachims 1997) and the ETL character dataset.
Dataset Splits No The paper mentions "training instances" and "testing data" but does not provide specific percentages, counts, or references to predefined train/validation/test dataset splits.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper mentions using a "Gaussian kernel" but does not specify any software dependencies with version numbers (e.g., library or solver names with versions).
Experiment Setup Yes The reduced set size is M = 10 which is very tiny size. As for our method, we set ν = 0.25 for the MPE estimator and choose the square loss ψ(z) = (1 z)2/4 for the job selector.