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 Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Rethinking Joint Maximum Mean Discrepancy for Visual Domain Adaptation
Authors: Wei Wang, Haifeng Xia, Chao Huang, Zhengming Ding, Cong Wang, Haojie Li, Xiaochun Cao
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on several cross-domain datasets could demonstrate the validity of our revealed theoretical results and the effectiveness of our proposed JMMD-HSIC. 4 Experiments 4.1 Datasets and Experimental Settings 4.2 Results 4.3 Feature Visualization 4.4 Ablation Study |
| Researcher Affiliation | Academia | 1Shenzhen Campus of Sun Yat-sen University 2Department of Computer Science, Tulane University 3University of California, San Francisco 4Shandong University of Science and Technology |
| Pseudocode | No | The paper describes methodologies using mathematical equations and descriptive text, but it does not include explicitly labeled pseudocode or algorithm blocks. The supplementary material provides mathematical derivations rather than structured algorithmic steps. |
| Open Source Code | Yes | Question: Does the paper provide open access to the data and code, with sufficient instructions to faithfully reproduce the main experimental results, as described in supplemental material? Answer: [Yes] Justification: We provided the references for the data and code. ... https://github.com/hellowangqian/domainadaptation-capls |
| Open Datasets | Yes | To validate our revealed theoretical results and the effectiveness of JMMD-HSIC, we conduct extensive experiments on four benchmark datasets in cross-domain object recognition. D1: Office10Caltech10 [47]... D2: Image CLEF-DA... D3: Office-31 [48]... D4: Office-Home [49] |
| Dataset Splits | No | The paper mentions several benchmark datasets (Office10-Caltech10, Image CLEF-DA, Office-31, Office-Home) and uses different features (SURF, DECAF-6, ResNet-50), but it does not explicitly provide specific percentages, sample counts, or citations for the train/test/validation splits used for these datasets in the main text. |
| Hardware Specification | No | Question: For each experiment, does the paper provide sufficient information on the computer resources (type of compute workers, memory, time of execution) needed to reproduce the experiments? Answer: [No] Justification: Our experiments do not require significant computational resources. |
| Software Dependencies | No | The paper mentions using specific features like SURF, DECAF-6, and ResNet-50, and classifiers like 1-NN, SVM, LP, and NCP. It also references libsvm in a footnote. However, it does not provide specific version numbers for any software components, libraries, or programming languages used. |
| Experiment Setup | Yes | Regarding δ, we uniformly set it to 0.5 for JDA+JMMD-HSIC, while assigning different values on the corresponding datasets for the other two variants after trials. ... Moreover, we adopt K2 for JMMD due to its superiority and K4 for HSIC. ... On Office10-Caltech10, we use the SURF features with 800 dimensions [47] and the DECAF-6 features with 4096 dimensions [52]. On the other three datasets, we utilize the Res Net-50 features with 2048 dimensions [53]. |