Parameter-Efficient Transfer from Sequential Behaviors for User Modeling and Recommendation

By Fajie Yuan, Xiangnan He, Alexandros Karatzoglou, Liguang Zhang. 2020

202 citations recorded by the source corpus.

Inductive transfer learning has had a big impact on computer vision and NLP domains but has not been used in the area of recommender systems. Even though there has been a large body of research on generating recommendations based on modeling user-item interaction sequences, few of them attempt to represent and transfer these models for serving downstream tasks where only limited data exists. In this paper, we delve on the task of effectively learning a single user representation that can be applied to a diversity of tasks, from cross-domain recommendations to user profile predictions. Fine-tuning a large pre-trained network and adapting it to downstream tasks is an effective way to solve such tasks. However, fine-tuning is parameter inefficient considering that an entire model needs to be re-trained for every new task. To overcome this issue, we develop a parameter-efficient transfer learning architecture, termed as PeterRec, which can be configured on-the-fly to various downstream tasks. Specifically, PeterRec allows the pre-trained parameters to remain unaltered during fine-tuning by injecting a series of re-learned neural networks, which are small but as expressive as learning the entire network. We perform extensive experimental ablation to show the effectiveness of the learned user representation in five downstream tasks. Moreover, we show that PeterRec performs efficient transfer learning in multiple domains, where it achieves comparable or sometimes better performance relative to fine-tuning the entire model parameters. Codes and datasets are available at https://github.com/fajieyuan/sigir2020_peterrec.

Subfields: Artificial Intelligence, Information Systems, Management Science and Operations Research

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Xiangnan He
2. Xiang Wang
3. Hongzhi Yin
4. Wayne Xin Zhao
5. Fuli Feng
6. Junliang Yu
7. Meng Wang
8. Tong Chen

In Information Systems

1. Xiangnan He
2. Xiang Wang
3. Hongzhi Yin
4. Junliang Yu
5. Fuli Feng
6. Meng Wang
7. Wayne Xin Zhao
8. Tong Chen

In Management Science and Operations Research

1. Xiangnan He
2. Xiang Wang
3. Meng Wang
4. Julian McAuley
5. Le Wu
6. Fuli Feng
7. Kun Zhang
8. Fajie Yuan

Citations in the corpus

• A Survey on Accuracy-Oriented Neural Recommendation: From Collaborative Filtering to Information-Rich Recommendation (2021, methodology): “Sequential Models RNN based models that rely on sessions to construct input and output p-RNN [202],KERL [193],CRNNs [213], NARM [214], [215], [216], [217], [201], Translation based models for modeling the correlations of consecutive items TransRec [46],PeterRec [218], [219]”
• Self-Supervised Learning for Recommender Systems: A Survey (2022, background): “There is also a line of research whose goal is to learn general-purpose representation through generative pre-training [125], [119], [120] so as to benefit multiple downstream recommendation tasks.”
• Where to Go Next for Recommender Systems? ID- vs. Modality-based Recommender Models Revisited (2023, background): “First, IDRec highly relies on the ID interactions, which fails to provide recommendations when users and items have few interactions [77, 79], a.”
• LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation (2020, methodology): “To alleviate information overload on the web, recommender system has been widely deployed to perform personalized information filtering [7, 45, 46].”
• TALLRec: An Effective and Efficient Tuning Framework to Align Large Language Model with Recommendation (2023, background): “Thus, some works focus on the generalization ability of sequential recommendation models by pretraining [32, 60], data augmentation [37, 38, 50, 56],”
• Iterative Network Pruning with Uncertainty Regularization for Lifelong Sentiment Classification (2021, methodology): “4The idea here is a bit similar to that used in [40], which however designs a different PRF and solves a very different research problem.”
• Large Language Models are Zero-Shot Rankers for Recommender Systems (2023, background): “We would like to take a more focused study on how they improve the recommendation performance of LLMs. Instruction tuning improves the ranking abilities of LLMs. Existing works show that instruction tuning significantly improves the generalization abilities of LLMs on unseen tasks [27, 32].”
• One Person, One Model, One World: Learning Continual User Representation without Forgetting (2020, methodology): “Particularly, in this paper we focus more on the finetuning strategy, which in general includes four typical ways [24, 39, 42].”