Enhancing ID-based Recommendation with Large Language Models

TL;DR

This work introduces a pioneering approach called “LLM for ID-based recommendation” (LLM4IDRec), which integrates the capabilities of LLMs while exclusively relying on ID data, thus diverging from the previous reliance on textual data.

Abstract

Large Language Models (LLMs) have recently garnered significant attention in various domains, including recommendation systems. Recent research leverages the capabilities of LLMs to improve the performance and user modeling aspects of recommender systems. These studies primarily focus on utilizing LLMs to interpret textual data in recommendation tasks. However, it's worth noting that in ID-based recommendations, textual data is absent, and only ID data is available. The untapped potential of LLMs for ID data within the ID-based recommendation paradigm remains relatively unexplored. To this end, we introduce a pioneering approach called "LLM for ID-based Recommendation" (LLM4IDRec). This innovative approach integrates the capabilities of LLMs while exclusively relying on ID data, thus diverging from the previous reliance on textual data. The basic idea of LLM4IDRec is that by employing LLM to augment ID data, if augmented ID data can improve recommendation performance, it demonstrates the ability of LLM to interpret ID data effectively, exploring an innovative way for the integration of LLM in ID-based recommendation. We evaluate the effectiveness of our LLM4IDRec approach using three widely-used datasets. Our results demonstrate a notable improvement in recommendation performance, with our approach consistently outperforming existing methods in ID-based recommendation by solely augmenting input data.

Figures (11)

• Figure 1: An example of comparing different structures of LLM in Recommender Systems (RS). Figure 1(a) and (b) depict current approaches for incorporating LLM into recommendation models. Both methods utilize textual data as input, such as user and item profiles. Nevertheless, in the context of ID-based recommendations, solely ID data is available, devoid of any textual information. When dealing with recommendation scenarios that solely rely on ID data, we investigated the application of LLM for ID-based recommendation, as shown in Figure 1(c).
• Figure 2: The architecture of our proposed LLM4IDRec.
• Figure 3: An example of generating data based on user sequence $S_u$ and prompt template.
• Figure 4: Analysis of augmented data $R_{aug}$ composition on Yelp data. Augmented data $R_{aug}$ consists of the original data $R$ and the generated data $R_{LLM}$ by LLM4IDRec. The generated data $R_{LLM}$ only accounts for 0.88% of the total data $R_{aug}$. Only 23.06% of the generated data $R_{LLM}$ coincides with the test data on Yelp data.
• Figure 5: Analysis of augmented data $R_{aug}$ composition on Amazon-kindle data. Augmented data $R_{aug}$ consists of the original data $R$ and the generated data $R_{LLM}$ by LLM4IDRec. The generated data $R_{LLM}$ only accounts for 0.48% of the total data $R_{aug}$. Only 22.93% of the generated data $R_{LLM}$ coincides with the test data on Amazon-kindle data.

Theorems & Definitions (4)

• definition 1
• Example 4.1
• Example 4.2
• Example 4.3