In-context Prompt-augmented Micro-video Popularity Prediction
Authors: Zhangtao Cheng, Jiao Li, Jian Lang, Ting Zhong, Fan Zhou
AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments conducted on three real-world datasets demonstrate the superiority of ICPF compared to 14 competitive baselines. |
| Researcher Affiliation | Academia | Zhangtao Cheng, Jiao Li, Jian Lang, Ting Zhong, Fan Zhou* University of Electronic Science and Technology of China, Chengdu, Sichuan, China EMAIL, jiao EMAIL, jian EMAIL, EMAIL, EMAIL |
| Pseudocode | No | The paper describes its methodology through textual descriptions and diagrams (e.g., Figure 2) but does not include explicit pseudocode or algorithm blocks. |
| Open Source Code | Yes | The source codes and datasets are available at https://github.com/Jolieresearch/ICPF. |
| Open Datasets | Yes | To analyze the effectiveness of our ICPF, we select three real-world micro-video datasets: Micro Lens (Ni et al. 2023), NUS (Chen et al. 2016), and Tik Tok (https://www.tiktok.com/), from various online video platforms. The source codes and datasets are available at https://github.com/Jolieresearch/ICPF. |
| Dataset Splits | Yes | Each dataset is randomly divided into training, validation, and test sets in a ratio of 8:1:1. |
| Hardware Specification | No | The paper does not provide specific hardware details such as GPU models, CPU types, or memory specifications used for running the experiments. |
| Software Dependencies | No | During retrieval, we utilize Vi TB/32 CLIP (Radford et al. 2021) as the image encoder and Angl E (Li and Li 2023) as the text encoder. While specific models/libraries are mentioned, their version numbers are not provided, nor are general software environments like Python or PyTorch with their versions. |
| Experiment Setup | Yes | We utilize the Adam W optimizer (Loshchilov and Hutter 2017) with a learning rate of 1 10 4 for optimizing the parameters. The model is trained for 30 epochs with a batch size of 64 and tested with a batch size of 256. |