CollageNoter: Real-Time and Adaptive Collage Layout Design for Screenshot-Based E-Note-Taking
Authors: Qiuyun Zhang, Bin Guo, Lina Yao, Xiaotian Qiao, Ying Zhang, Zhiwen Yu
AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experimental results have confirmed the effectiveness of our Collage Noter. Both qualitative and quantitative experiments validate the effectiveness of our approach compared to several strong baselines. Moreover, we conduct user studies to demonstrate the advantages our methods provide to users. |
| Researcher Affiliation | Academia | 1 School of Computer Science, Northwestern Polytechnical University 2 CSIRO s Data61 3 School of Computer Science and Technology, Xidian University 4 Guangzhou Institute of Technology, Xidian University 5 College of Computer Science and Technology, Harbin Engineering University EMAIL, EMAIL, EMAIL, EMAIL, EMAIL, EMAIL |
| Pseudocode | No | The paper describes methods using text and equations but does not include any explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide any specific links to a code repository, an explicit statement about code release, or mention code in supplementary materials. |
| Open Datasets | Yes | Many public datasets related to different graphic layout design tasks have been proposed including Magazine (Zheng et al. 2019) for magazine, Rico (Deka et al. 2017) for user interface, Publey Net (Zhong, Tang, and Yepes 2019) for document, and PKU (Zhou et al. 2022) for advertisements, etc. |
| Dataset Splits | No | The paper mentions training data and a collected dataset but does not specify exact training, validation, or test splits (e.g., percentages or counts) for any of the datasets used. |
| Hardware Specification | Yes | All experiments are carried out with Pytorch framework and NVIDIA 3080 Ti GPUs. |
| Software Dependencies | No | The paper mentions using the "Pytorch framework" but does not specify a version number or other software dependencies with their versions. |
| Experiment Setup | Yes | Each transformer block has 4 layers with a hidden size of 512, 4-head attention, and feed-forward dim of 1028. The Collage Former is trained for 100 epochs. The Adam optimizers are used, and the initial learning rate is 10 3 for both generator and discriminator. |