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 Shift Toward Open and Reproducible AI Research," Under Review, 2026. doi: 10.48550/arXiv.2606.16974..
Multimodal Negative Learning
Authors: Baoquan Gong, Xiyuan Gao, Pengfei Zhu, Qinghua Hu, Bing Cao
NeurIPS 2025 | Venue PDF | LLM Run Details | Input Tokens: 31,275 Total number of tokens sent to the LLM as input for this paper's analysis. | Output Tokens: 3,436 Total number of tokens produced by the LLM (including reasoning/thinking tokens) for this paper's analysis.
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments across multiple benchmarks demonstrate the effectiveness and generalizability of our approach against the competing methods. |
| Researcher Affiliation | Academia | Baoquan Gong Xiyuan Gao Pengfei Zhu Qinghua Hu Bing Cao School of Artificial Intelligence, Tianjin University, Tianjin, China Haihe Lab of ITAI EMAIL |
| Pseudocode | Yes | The overall pseudocode of our model is provided in the Appendix C.6. |
| Open Source Code | Yes | The code is available at https: //github.com/Baoquan Gong/Multimodal-Negative-Learning.git. |
| Open Datasets | Yes | All datasets used in this work are publicly available. |
| Dataset Splits | Yes | For data splits, hyperparameter selection, and the type of optimizer used, we follow the same settings as the baselines. |
| Hardware Specification | Yes | All experiments are conducted on an NVIDIA TITAN GPU using Py Torch, with default settings applied across all methods. |
| Software Dependencies | No | All experiments are conducted on an NVIDIA TITAN GPU using Py Torch, with default settings applied across all methods. |
| Experiment Setup | Yes | All models are trained for 100 epochs with a batch size of 16, using the Adam optimizer. The learning rate is set to 1e-4, with a warmup proportion of 0.1. |