Mixup-Induced Domain Extrapolation for Domain Generalization

Authors: Meng Cao, Songcan Chen

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In experiments, EDM has been plugged into several methods in both closed and open set settings, achieving up to 5.73% improvement. Extensively experiments are constructed to comprehensively evaluate the effectiveness of EDM on two datasets both in closed and open set settings.
Researcher Affiliation Academia Meng Cao1,2, Songcan Chen1,2* 1College of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics 2MIIT Key Laboratory of Pattern Analysis and Machine Intelligence {meng.cao, s.chen}@nuaa.edu.cn
Pseudocode Yes Detailed Algorithm is provided in Appendix.
Open Source Code Yes The code is available at https://github.com/Alrash/EDM.
Open Datasets Yes For the architecture, we use Res Net-18 as backbone on three datasets, i.e., PACS, Office-Home, and Domain Net datasets.
Dataset Splits Yes For both settings, we follow corresponding settings from the previous methods, i.e., the same closed-set setting as (Lu et al. 2022), and the same open-set setting as (Shu et al. 2021).
Hardware Specification No The paper does not provide specific details about the hardware used for experiments, such as GPU models, CPU types, or memory.
Software Dependencies No The paper does not specify version numbers for any software dependencies or libraries used in the experiments.
Experiment Setup No The paper states 'For both settings, we follow corresponding settings from the previous methods,' but does not provide specific hyperparameter values or detailed training configurations within the main text.