SepFusion: Finding Optimal Fusion Structures for Visual Sound Separation
Authors: Dongzhan Zhou, Xinchi Zhou, Di Hu, Hang Zhou, Lei Bai, Ziwei Liu, Wanli Ouyang3544-3552
AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we first introduce the datasets and evaluation metrics. The searching and training details will be provided in the supplementary. Then we compare the performance of Sep Fusion with the baseline methods and show the ablation studies. |
| Researcher Affiliation | Collaboration | Dongzhan Zhou1*, Xinchi Zhou1*, Di Hu2 , Hang Zhou3, Lei Bai1, Ziwei Liu4, Wanli Ouyang1 1The University of Sydney, 2Gaoling School of Artificial Intelligence, Renmin University of China, 3Baidu Inc., 4S-lab, Nanyang Technological University |
| Pseudocode | Yes | The generation steps of each intermediate node are as follows, and the selecting action in each step is random selection. Step 1. Select two inputs from the existing node set. Step 2. Select an operation from the operation pool. Step 3. Apply the operation on the two input nodes to compute the non-activated feature. Step 4. Select one activation function from sigmoid, softmax, Re LU, and identity (no activation). Activate the feature from Step 3 to produce the new node. Step 5. Add the new node to the node set. |
| Open Source Code | No | No explicit statement or link providing concrete access to the source code for the methodology described in this paper. |
| Open Datasets | Yes | We validate the effects of our fusion module on the MUSIC dataset (Zhao et al. 2019), which contains 21 classes of instruments. |
| Dataset Splits | Yes | We validate the effects of our fusion module on the MUSIC dataset (Zhao et al. 2019)... The searching and training details will be provided in the supplementary. We conduct experiments on the Fair-Play dataset and the results are listed in Table 3. |
| Hardware Specification | No | No specific hardware details (e.g., exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) are provided for running experiments. |
| Software Dependencies | No | The paper mentions 'open-source mir eval library (Raffel et al. 2014)' but does not provide specific version numbers for this or any other software dependencies. |
| Experiment Setup | Yes | The searching and training details will be provided in the supplementary. |