Probabilistic Multimodal Learning with von Mises-Fisher Distributions

Authors: Peng Hu, Yang Qin, Yuanbiao Gou, Yunfan Li, Mouxing Yang, Xi Peng

IJCAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on nine benchmarks demonstrate the superiority of PML, consistently outperforming 14 state-of-the-art methods. Code is available at https://github. com/XLearning-SCU/2025-IJCAI-PML. The paper includes a dedicated section 4 titled 'Experiments' with subsections for 'Comparisons with State of the Art', 'Ablation Study', and 'Reliability Study', presenting numerous performance tables and analysis.
Researcher Affiliation Academia 1College of Computer Science, Sichuan University, China. 2National Key Laboratory of Fundamental Algorithms and Models for Engineering Simulation, China. All listed affiliations are academic institutions in China.
Pseudocode No The paper describes its methodology using mathematical formulations and descriptive text within the 'Methodology' section, but it does not include any explicitly labeled pseudocode or algorithm blocks.
Open Source Code Yes Code is available at https://github. com/XLearning-SCU/2025-IJCAI-PML.
Open Datasets Yes We evaluate our framework on nine publicly available benchmark datasets spanning diverse modalities... These datasets comprise Handwritten[Duin, 1998], MSRC-V11, NUSWIDE-OBJECT2 (NUSOBJ) [Chua et al., 2009], Fashion MV [Xu et al., 2022], Scene153, Land Use [Yang and Newsam, 2010], Leaves1004, MVSA [Niu et al., 2016] and UPMC-Food101 [Wang et al., 2015].
Dataset Splits No We report the accuracies on the test set to measure the performance. In our experiments, in addition to the normal experimental setting, we also construct a noise setting following [Xu et al., 2024a] to further evaluate the robustness of our method by introducing data noise and noisy correspondences (misaligned views/modalities) on test sets. The paper mentions using 'test sets' but does not specify exact percentages, sample counts, or a detailed splitting methodology for training, validation, and test sets.
Hardware Specification No The paper does not provide specific details about the hardware used for running experiments, such as GPU models, CPU types, or memory specifications.
Software Dependencies No The paper mentions using the 'Adam optimizer' and designing 'multi-layer linear networks' for sub-networks, but it does not specify versions of any software libraries, frameworks (e.g., PyTorch, TensorFlow), or programming languages used for implementation.
Experiment Setup Yes In our experiments, the sub-networks fv and gv, which estimate the mean direction and concentration parameter, respectively, are designed as multi-layer linear networks. We exploit the Adam optimizer with a batch size of 32 to train all models, using a learning rate of 1e-4 for all datasets.