Kumaraswamy Wavelet for Heterophilic Scene Graph Generation
Authors: Lianggangxu Chen, Youqi Song, Shaohui Lin, Changbo Wang, Gaoqi He
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Comprehensive experiments on the Visual Genome and Open Images datasets show that our method achieves state-of-the-art performance. |
| Researcher Affiliation | Academia | 1School of Computer Science and Technology, East China Normal University, Shanghai, China 2Chongqing Key Laboratory of Precision Optics, Chongqing Institute of East China Normal University, Chongqing, China |
| Pseudocode | No | The paper describes the proposed method using text and mathematical equations but does not include any explicitly labeled 'Pseudocode' or 'Algorithm' blocks. |
| Open Source Code | No | The paper does not contain any explicit statement about releasing the source code for the described methodology or a link to a code repository. |
| Open Datasets | Yes | We use the large-scale VG dataset (Krishna et al. 2017) to evaluate the effectiveness of the proposed KWGNN. |
| Dataset Splits | Yes | The training set of VG includes 70% images, with 5K images used as a validation subset. The testing set is composed of the remaining 30% of the images. |
| Hardware Specification | No | The paper mentions training time ('approximately 19-20 hours') but does not provide specific details about the hardware used, such as GPU/CPU models or memory. |
| Software Dependencies | No | The paper mentions using the Adam optimizer but does not specify any software dependencies with version numbers (e.g., Python, PyTorch, TensorFlow versions). |
| Experiment Setup | Yes | The Adam (Kingma and Ba 2014) optimizer is used with batch size 4 during training. The initial learning rate is 0.0001 for the backbone and 0.0003 for other parts, which decays by a factor of 10 for every 5 epochs. The weight decay is set as 0.0001. We set the dimension of node representations to 1024, and the filter number C in KWGNN is 3. The α and γ are empirically set to 0.1 and 2, respectively. |