Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

IDGenRec: LLM-RecSys Alignment with Textual ID Learning

Juntao Tan, Shuyuan Xu, Wenyue Hua, Yingqiang Ge, Zelong Li, Yongfeng Zhang

TL;DR

This work tackles the item-encoding bottleneck in generative recommender systems by learning meaningful textual IDs for items via an ID generator, enabling seamless text-to-text integration with LLM-based recommenders. It introduces Diverse ID Generation to ensure concise yet unique IDs and an alternating training scheme to align the ID generator with the base recommender. Empirically, IDGenRec achieves strong improvements over baselines in standard sequential recommendation and demonstrates promising zero-shot performance when trained on a large, cross-domain corpus, suggesting potential as a foundation model for generative recommendations. The approach leverages a T5-based backbone for both components and provides open-source code and data for replication.

Abstract

Generative recommendation based on Large Language Models (LLMs) have transformed the traditional ranking-based recommendation style into a text-to-text generation paradigm. However, in contrast to standard NLP tasks that inherently operate on human vocabulary, current research in generative recommendations struggles to effectively encode recommendation items within the text-to-text framework using concise yet meaningful ID representations. To better align LLMs with recommendation needs, we propose IDGen, representing each item as a unique, concise, semantically rich, platform-agnostic textual ID using human language tokens. This is achieved by training a textual ID generator alongside the LLM-based recommender, enabling seamless integration of personalized recommendations into natural language generation. Notably, as user history is expressed in natural language and decoupled from the original dataset, our approach suggests the potential for a foundational generative recommendation model. Experiments show that our framework consistently surpasses existing models in sequential recommendation under standard experimental setting. Then, we explore the possibility of training a foundation recommendation model with the proposed method on data collected from 19 different datasets and tested its recommendation performance on 6 unseen datasets across different platforms under a completely zero-shot setting. The results show that the zero-shot performance of the pre-trained foundation model is comparable to or even better than some traditional recommendation models based on supervised training, showing the potential of the IDGen paradigm serving as the foundation model for generative recommendation. Code and data are open-sourced at https://github.com/agiresearch/IDGenRec.

IDGenRec: LLM-RecSys Alignment with Textual ID Learning

TL;DR

This work tackles the item-encoding bottleneck in generative recommender systems by learning meaningful textual IDs for items via an ID generator, enabling seamless text-to-text integration with LLM-based recommenders. It introduces Diverse ID Generation to ensure concise yet unique IDs and an alternating training scheme to align the ID generator with the base recommender. Empirically, IDGenRec achieves strong improvements over baselines in standard sequential recommendation and demonstrates promising zero-shot performance when trained on a large, cross-domain corpus, suggesting potential as a foundation model for generative recommendations. The approach leverages a T5-based backbone for both components and provides open-source code and data for replication.

Abstract

Generative recommendation based on Large Language Models (LLMs) have transformed the traditional ranking-based recommendation style into a text-to-text generation paradigm. However, in contrast to standard NLP tasks that inherently operate on human vocabulary, current research in generative recommendations struggles to effectively encode recommendation items within the text-to-text framework using concise yet meaningful ID representations. To better align LLMs with recommendation needs, we propose IDGen, representing each item as a unique, concise, semantically rich, platform-agnostic textual ID using human language tokens. This is achieved by training a textual ID generator alongside the LLM-based recommender, enabling seamless integration of personalized recommendations into natural language generation. Notably, as user history is expressed in natural language and decoupled from the original dataset, our approach suggests the potential for a foundational generative recommendation model. Experiments show that our framework consistently surpasses existing models in sequential recommendation under standard experimental setting. Then, we explore the possibility of training a foundation recommendation model with the proposed method on data collected from 19 different datasets and tested its recommendation performance on 6 unseen datasets across different platforms under a completely zero-shot setting. The results show that the zero-shot performance of the pre-trained foundation model is comparable to or even better than some traditional recommendation models based on supervised training, showing the potential of the IDGen paradigm serving as the foundation model for generative recommendation. Code and data are open-sourced at https://github.com/agiresearch/IDGenRec.
Paper Structure (24 sections, 5 equations, 3 figures, 7 tables, 1 algorithm)

This paper contains 24 sections, 5 equations, 3 figures, 7 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Approach
  3. Generative Process
  4. ID Generator
  5. Diverse ID Generation
  6. Base Recommender
  7. Alternate Training
  8. Training Base Recommender
  9. Training ID Generator
  10. Model Initialization
  11. Experiment
  12. Datasets
  13. Implementation Details
  14. Evaluation Metrics
  15. Exp1: Standard Evaluation
  16. ...and 9 more sections

Figures (3)

  • Figure 1: The ID generator takes plain text from each item's meta textual information and generates abstractive textual IDs for the item's representation.
  • Figure 2: A real example showing the generative recommendation workflow. The ID generator generates item IDs for items from the user history by taking their plain text. Then, the generated IDs are interpolated into the template. Addtionally, the user ID is generated by using all items' text in the user's history, showing a "high-level profile" of the user's preference. The position embeddings are subsequently combined with token embeddings to capture the sequence of interactions. Finally, the base recommender generates the ID of the recommended item based on constrained decoding.
  • Figure 3: Quality study of the generated item IDs: two examples for each dataset, one with lengthy plain text data and one with shorter plain text data. The blue IDs are generated by the initial ID generator pre-trained on article tag generation, while the green IDs are generated by the ID generator after alternate training.