- LPProgressive Layered Extraction (PLE): A Novel Multi-Task Learning (MTL) Model for Personalized Recommendations
by Hongyan Tang (Tencent PCG), Junning Liu (Tencent PCG), Ming Zhao (Tencent PCG), Xudong Gong (Tencent PCG)
Multi-task learning (MTL) has been successfully applied to many recommendation applications. However, MTL models often suffer from performance degeneration with negative transfer due to the complex and competing task correlation in real-world recommender systems. Moreover, through extensive experiments across SOTA MTL models, we have observed an interesting seesaw phenomenon that performance of one task is often improved by hurting the performance of some other tasks. To address these issues, we propose a Progressive Layered Extraction (PLE) model with a novel sharing structure design. PLE separates shared components and task-specific components explicitly and adopts a progressive routing mechanism to extract and separate deeper semantic knowledge gradually, improving efficiency of joint representation learning and information routing across tasks in a general setup. We apply PLE to both complicatedly correlated and normally correlated tasks, ranging from two-task cases to multi-task cases on a real-world Tencent video recommendation dataset with 1 billion samples, and results show that PLE outperforms state-of-the-art MTL models significantly under different task correlations and task-group size. Furthermore, online evaluation of PLE on a large-scale content recommendation platform at Tencent manifests 2.23% increase in view-count and 1.84% increase in watch time compared to SOTA MTL models, which is a significant improvement and demonstrates the effectiveness of PLE. Finally, extensive offline experiments on public benchmark datasets demonstrate that PLE can be applied to a variety of scenarios besides recommendations to eliminate the seesaw phenomenon. PLE now has been deployed to the online video recommender system in Tencent successfully.
Full text in ACM Digital Library
|
- LPKRED: Knowledge-Aware Document Representation for News Recommendations
by Danyang Liu (University of Science and Technologyof China), Jianxun Lian (Microsoft Research Asia), Shiyin Wang (Tsinghua University), Ying Qiao (Microsoft Corp.), Jiun-Hung Chen (Microsoft Corp.), Guangzhong Sun (University of Science and Technology of China), Xing Xie (Microsoft Research Asia)
News articles usually contain knowledge entities such as celebrities or organizations. Important entities in articles carry key messages and help to understand the content in a more direct way. An industrial news recommender system contains various key applications, such as personalized recommendation, item-to-item recommendation, news category classification, news popularity prediction and local news detection. We find that incorporating knowledge entities for better document understanding benefits these applications consistently. However, existing document understanding models either represent news articles without considering knowledge entities (e.g., BERT) or rely on a specific type of text encoding model (e.g., DKN) so that the generalization ability and efficiency is compromised. In this paper, we propose KRED, which is a fast and effective model to enhance arbitrary document representation with a knowledge graph. KRED first enriches entities’ embeddings by attentively aggregating information from their neighborhood in the knowledge graph. Then a context embedding layer is applied to annotate the dynamic context of different entities such as frequency, category and position. Finally, an information distillation layer aggregates the entity embeddings under the guidance of the original document representation and transforms the document vector into a new one. We advocate to optimize the model with a multi-task framework, so that different news recommendation applications can be united and useful information can be shared across different tasks. Experiments on a real-world Microsoft News dataset demonstrate that KRED greatly benefits a variety of news recommendation applications.
Full text in ACM Digital Library
|
|
- LPFISSA: Fusing Item Similarity Models with Self-Attention Networks for Sequential Recommendation
by Jing Lin (College of Computer Science and Software Engineering, Shenzhen University), Weike Pan (College of Computer Science and Software Engineering, Shenzhen University), Zhong Ming (College of Computer Science and Software Engineering, Shenzhen University)
Sequential recommendation has been a hot research topic because of its practicability and high accuracy by capturing the sequential information. As deep learning (DL) based methods being widely adopted to model the local and dynamic preferences beneath users’ behavior sequences, the modeling of users’ global and static preferences tends to be underestimated that usually, only some simple and crude users’ latent representations are introduced. Moreover, most existing methods hold an assumption that users’ intention can be fully captured by considering the historical behaviors, while neglect the possible uncertainty of users’ intention in reality, which may be influenced by the appearance of the candidate items to be recommended. In this paper, we thus focus on these two issues, i.e., the imperfect modeling of users’ global preferences in most DL-based sequential recommendation methods and the uncertainty of users’ intention brought by the candidate items, and propose a novel solution named fusing item similarity models with self-attention networks (FISSA) for sequential recommendation. Specifically, we treat the state-of-the-art self-attentive sequential recommendation (SASRec) model as the local representation learning module to capture the dynamic preferences beneath users’ behavior sequences in our FISSA, and further propose a global representation learning module to improve the modeling of users’ global preferences and a gating module that balances the local and global representations by taking the information of the candidate items into account. The global representation learning module can be seen as a location-based attention layer, which is effective to fit in well with the parallelization training process of the self-attention framework. The gating module calculates the weight by modeling the relationship among the candidate item, the recently interacted item and the global preference of each user using an MLP layer. Extensive empirical studies on five commonly used datasets show that our FISSA significantly outperforms eight state-of-the-art baselines in terms of two commonly used metrics.
Full text in ACM Digital Library
|
- INInvestigating Multimodal Features for Video Recommendations at Globoplay
by Felipe Ferreira (Pontifical Catholic University of Rio de Janeiro, Globo.com), Daniele R. Souza (Globo.com), Igor Moura (Globo.com), Matheus Barbieri (Globo.com), Hélio C. V. Lopes (Pontifical Catholic University of Rio de Janeiro)
“Globoplay is Globo Group’s digital video streaming platform and offers a very diverse video content catalogue ranging from international to brazilian productions such as movies, series, soap operas, and TV programs produced by Globo Group. One of the challenges with such large and diverse content collection is its distribution to the user base in order to help our subscribers with finding relevant content that meets their expectations and to increase their engagement with the product. In this work, we show the result of a content-based recommendation approach based on multi-modal features such as visual characteristics and audio patterns found in the video content. Using techniques applied to short videos, we model it as a similarity problem based on the content of the video, where, given a video, we establish the top-n videos most similar to it in the collection. For the evaluation, we conducted a study through interviews with a group of users to understand their perception of recommendations based on audiovisual characteristics. For the future, we plan to: explore and define the best approach to combine text, audio and video features for video recommendations; explore audiovisual features with other recommendation approaches such as session based and collaborative filtering; perform AB testing in production; and evaluate the proposal impact in business metrics.”
Full text in ACM Digital Library
|
|