用内容驱动的压缩模块，高效生成微视频推荐嵌入。
Compressed Video Aggregator: Content-driven Module for Efficient Micro-Video Recommendation
- 通过掩码均值池化提取帧嵌入语义锚点，降低冗余。
- 训练时间与显存消耗减少数量级，性能提升显著。
- 适用于对效率敏感的微视频推荐场景，抗标题错误性强。
我们提出轻量级微视频推荐模块 Compressed Video Aggregator (CVA)，将视频信息与偏好学习解耦。CVA 首先利用掩码均值池化，将冻结的视频帧嵌入（VFM）压缩为语义一致的锚点，再投影至紧凑潜在空间，并通过残差自注意力和前馈网络优化表示，最终生成单一视频嵌入供推荐系统使用。针对原始基准数据集中帧数冗余及采样过粗的问题，我们基于 CLIP 利用标题重选关键帧。在 MicroLens 与 Short-Video 数据集上的实验表明，该方法在训练时间与 GPU 显存消耗上实现数量级降低，且性能持续提升；重选帧进一步增强了所有方法的表现，包括 CVA 本身。此外，我们还分析了错误标题对方法的影响。
原文摘要 · Abstract (English)
We propose \textbf{Compressed Video Aggregator} (CVA), a lightweight micro-video recommendation module that decouples video information from preference learning. CVA first summarizes frozen VFM frame embeddings into a semantic-consensus anchor through masked mean pooling, projects this anchor into a compact latent space, and refines the projected representation with residual self-attention and feedforward blocks before producing a single video embedding for the recommender. Due to the redundancy in the frame count of the original benchmark and its overly coarse sampling, we used titles to re-select key frames based on CLIP. Experiments on MicroLens and Short-Video show consistent gains with orders-of-magnitude reductions in training time and GPU memory, and re-selected frames can further enhance the performance of all methods, including CVA. Furthermore, we also discussed the impact of several scenarios involving erroneous titles on our method.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。