Deep Mutual Learning Optimizes Multi-Task Recommender Systems with Cross Task Feature Mining

This paper presents Deep Mutual Learning (DML), a framework for multi-task recommender systems. It improves ranking by introducing Cross Task Feature Mining (CTFM) to share task-specific inputs while ensuring task awareness. The design avoids negative task interactions, enhances predictions, and encourages effective knowledge transfer via scaled dot-product attention and task embeddings.


This content originally appeared on HackerNoon and was authored by KMeans

Abstract and 1 Introduction

  1. Methodology
  2. Experiments
  3. Conclusion and References

2. METHODOLOGY

In this section, we first introduce the problem of multi-objective ranking for recommender systems. Second, we describe the general design of DML. Finally, we elaborate on the introduced components.

2.1 Multi-Objective Ranking for Recsys

2.2 Overall Design of DML

Please refer to figure 1 for the overall network architecture of DML. With the existing MTL algorithms[25], the equation (1) can be further decomposed as below. For simplicity, we omit the subscript 𝑛 in this section.

\

\ In this research, rather than 𝐺, we focus on the enhancement of upper-level networks for improved prediction performance. First, the shared component of CTFM is introduced, which leverages the attention mechanism to extract relevant information from the inputs of other task towers (the results of Equation (3)) as a complement to the target task. Please note that this attention is welldesigned to solve the task-awareness missing issue, for which the excessive encouragement of knowledge sharing is not conducive to the extraction of task-specific knowledge. With our design, the gradients computed from the target task’s loss will not impact the inputs of other task towers.

\

2.3 Cross Task Feature Mining

\

\ Table 1: The overall performance. The bold-face font denotes the winner in that column. Moreover, the "*" symbol denotes introducing DML achieves significant (p < 0.05 for one-tailed t-test) gain over the corresponding baseline.

\

:::info Authors:

(1) Yi Ren, Tencent, Beijing, China (yiren_bj@outlook.com);

(2) Ying Du, Tencent, Beijing, China (yingdu@tencent.com);

(3) Bin Wang, Tencent, Beijing, China (hillmwang@tencent.com);

(4) Shenzheng Zhang, Tencent, Beijing, China (qjzcyzhang@tencent.com).

:::

:::info This paper is available on arxiv under CC BY 4.0 DEED license.

:::

\


This content originally appeared on HackerNoon and was authored by KMeans


Print Share Comment Cite Upload Translate Updates
APA

KMeans | Sciencx (2025-01-23T19:00:19+00:00) Deep Mutual Learning Optimizes Multi-Task Recommender Systems with Cross Task Feature Mining. Retrieved from https://www.scien.cx/2025/01/23/deep-mutual-learning-optimizes-multi-task-recommender-systems-with-cross-task-feature-mining/

MLA
" » Deep Mutual Learning Optimizes Multi-Task Recommender Systems with Cross Task Feature Mining." KMeans | Sciencx - Thursday January 23, 2025, https://www.scien.cx/2025/01/23/deep-mutual-learning-optimizes-multi-task-recommender-systems-with-cross-task-feature-mining/
HARVARD
KMeans | Sciencx Thursday January 23, 2025 » Deep Mutual Learning Optimizes Multi-Task Recommender Systems with Cross Task Feature Mining., viewed ,<https://www.scien.cx/2025/01/23/deep-mutual-learning-optimizes-multi-task-recommender-systems-with-cross-task-feature-mining/>
VANCOUVER
KMeans | Sciencx - » Deep Mutual Learning Optimizes Multi-Task Recommender Systems with Cross Task Feature Mining. [Internet]. [Accessed ]. Available from: https://www.scien.cx/2025/01/23/deep-mutual-learning-optimizes-multi-task-recommender-systems-with-cross-task-feature-mining/
CHICAGO
" » Deep Mutual Learning Optimizes Multi-Task Recommender Systems with Cross Task Feature Mining." KMeans | Sciencx - Accessed . https://www.scien.cx/2025/01/23/deep-mutual-learning-optimizes-multi-task-recommender-systems-with-cross-task-feature-mining/
IEEE
" » Deep Mutual Learning Optimizes Multi-Task Recommender Systems with Cross Task Feature Mining." KMeans | Sciencx [Online]. Available: https://www.scien.cx/2025/01/23/deep-mutual-learning-optimizes-multi-task-recommender-systems-with-cross-task-feature-mining/. [Accessed: ]
rf:citation
» Deep Mutual Learning Optimizes Multi-Task Recommender Systems with Cross Task Feature Mining | KMeans | Sciencx | https://www.scien.cx/2025/01/23/deep-mutual-learning-optimizes-multi-task-recommender-systems-with-cross-task-feature-mining/ |

Please log in to upload a file.



There are no updates yet.
Click the Upload button above to add an update.

You must be logged in to translate posts. Please log in or register.

Related Posts