BeFA: A General Behavior-driven Feature Adapter for Multimedia Recommendation

Authors: Qile Fan, Penghang Yu, Zhiyi Tan, Bing-Kun Bao, Guanming Lu

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

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments demonstrate the effectiveness of the adapter across all multimedia recommendation methods. We conducted experiments on three publicly available datasets: (a) TMALL; (b) Microlens; and (c) H&M. The detailed information is presented in the Appendix. Table 1: Performance Comparison on Different Recommender Models.
Researcher Affiliation Academia 1School of Communications and Information Engineering, Nanjing University of Posts and Telecommunications, Nanjing, China 2Jiangsu Key Laboratory of Intelligent Information Processing and Communication Technology, Nanjing, China 3School of Computer Science, Nanjing University of Posts and Telecommunications, Nanjing, China EMAIL, EMAIL, EMAIL, EMAIL, EMAIL
Pseudocode No The paper describes the Behavior-driven Feature Adapter (Be FA) and its components using mathematical formulas (Equations 5, 6, 7) and textual descriptions, but it does not include any clearly labeled pseudocode or algorithm blocks.
Open Source Code No The paper mentions implementing MMRec (a unified public repository for multimodal recommendation methods) and provides a link to its GitHub repository (https://github.com/enoche/MMRec). However, it does not explicitly state that the source code for the proposed Be FA methodology is released or included within this repository, making the statement ambiguous regarding the availability of their specific contribution's code.
Open Datasets Yes We conducted experiments on three publicly available datasets: (a) TMALL1; (b) Microlens2; and (c) H&M3. 1https://tianchi.aliyun.com/dataset/140281 2https://recsys.westlake.edu.cn/Micro Lens-50k-Dataset/ 3https://www.kaggle.com/datasets/odins0n/handm-dataset128x128
Dataset Splits Yes For a fair comparison, we follow the evaluation settings in (Zhang et al. 2021; Zhou et al. 2023c) with the same 8:1:1 data splitting strategy for training, validation and testing.
Hardware Specification Yes Our experiments are done using RTX 4090 on windows system.
Software Dependencies No The paper mentions implementing MMRec based on PyTorch, but it does not specify the version number of PyTorch or any other software dependencies with their versions.
Experiment Setup Yes For general settings, we initialized embeddings with Xavier initialization of dimension 64, set the regularization coefficient to λE = 10-4, and batch size to B = 2048. Early stopping and total epochs are fixed at 10 and 1000, respectively.