Treffer: 融合文本增强与深度强化学习的交互式推荐方法.

To address the data-sparsity bottleneck and the inefficiency caused by large discrete action spaces in interactive recommender systems(IRS), this paper proposed an interactive recommendation method that tightly integrated text enhancement with deep reinforcement learning (TDIRS). The method first leveraged rich textual information to generate expressive embeddings and built a deep reinforcement-learning model powered by multi-head attention to capture users’ evolving preferences, thereby mitigating sparsity. Simultaneously, the method designed a dynamic candidate-action generator that fused item popularity with embedding representations to identify items that truly interested the user and to shrink the action space, boosting recommendation efficiency. Extensive experiments on the music dataset and others confirm that TDIRS outperforms all baselines on core metrics, achieving up to a improvement of 7percentage points in HR@10 and validating the effectiveness of the approach. [ABSTRACT FROM AUTHOR]

针对交互式推荐系统 (IRS) 仍然存在深度强化学习 (DRL) 所带来的数据稀疏性和大型离散动作空间 导致的效率低下问题, 提出了一种融合文本增强与深度强化学习的交互式推荐方法 (TDIRS)。首先利用文本信 息来获得丰富的嵌入表示, 并构建一个基于多头注意力机制的深度强化学习模型以捕捉用户不断变化的偏好, 从而缓解稀疏问题; 同时, 给出一种基于项目流行度和嵌入表示的动态候选动作生成方法来识别用户感兴趣的 项目, 并缩小动作空间, 从而提高推荐性能。实验表明, TDIRS 在 music 等数据集上的核心指标优于其基线方法, 其中 HR@10 最高提升了 7 个百分点。实验结果验证了 TDIRS 的有效性。 [ABSTRACT FROM AUTHOR]

Copyright of Application Research of Computers / Jisuanji Yingyong Yanjiu is the property of Application Research of Computers Edition and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)