Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Adaptive In-Context Learning with Large Language Models for Bundle Generation

Zhu Sun, Kaidong Feng, Jie Yang, Xinghua Qu, Hui Fang, Yew-Soon Ong, Wenyuan Liu

TL;DR

This work tackles the challenge of generating fixed-size, semantically coherent bundles while extracting underlying user intents from sessions. It introduces Adaptive In-Context Learning (AICL), a framework that leverages Neighbor Session Retrieval (NSR), Adaptive Demonstration Generation (ADG), and Demonstration Guided Inference (DGI) to learn from related sessions via tailored demonstrations. AICL employs mutual self-correction and adaptive auto-feedback to improve reliability and reduce hallucinations, and it uses rules summarization to prevent repeated errors. Experiments on three real-world datasets show AICL outperforms baselines in bundle generation and yields high-quality intent inferences, demonstrating the potential of adaptive, demonstration-driven LLMs for complex, multi-task recommendation problems. The approach advances practical bundle generation by delivering multiple, personalized, and intelligible bundles with adaptive sizes aligned to user intents, with implications for downstream tasks and scalable deployment.

Abstract

Most existing bundle generation approaches fall short in generating fixed-size bundles. Furthermore, they often neglect the underlying user intents reflected by the bundles in the generation process, resulting in less intelligible bundles. This paper addresses these limitations through the exploration of two interrelated tasks, i.e., personalized bundle generation and the underlying intent inference, based on different user sessions. Inspired by the reasoning capabilities of large language models (LLMs), we propose an adaptive in-context learning paradigm, which allows LLMs to draw tailored lessons from related sessions as demonstrations, enhancing the performance on target sessions. Specifically, we first employ retrieval augmented generation to identify nearest neighbor sessions, and then carefully design prompts to guide LLMs in executing both tasks on these neighbor sessions. To tackle reliability and hallucination challenges, we further introduce (1) a self-correction strategy promoting mutual improvements of the two tasks without supervision signals and (2) an auto-feedback mechanism for adaptive supervision based on the distinct mistakes made by LLMs on different neighbor sessions. Thereby, the target session can gain customized lessons for improved performance by observing the demonstrations of its neighbor sessions. Experiments on three real-world datasets demonstrate the effectiveness of our proposed method.

Adaptive In-Context Learning with Large Language Models for Bundle Generation

TL;DR

This work tackles the challenge of generating fixed-size, semantically coherent bundles while extracting underlying user intents from sessions. It introduces Adaptive In-Context Learning (AICL), a framework that leverages Neighbor Session Retrieval (NSR), Adaptive Demonstration Generation (ADG), and Demonstration Guided Inference (DGI) to learn from related sessions via tailored demonstrations. AICL employs mutual self-correction and adaptive auto-feedback to improve reliability and reduce hallucinations, and it uses rules summarization to prevent repeated errors. Experiments on three real-world datasets show AICL outperforms baselines in bundle generation and yields high-quality intent inferences, demonstrating the potential of adaptive, demonstration-driven LLMs for complex, multi-task recommendation problems. The approach advances practical bundle generation by delivering multiple, personalized, and intelligible bundles with adaptive sizes aligned to user intents, with implications for downstream tasks and scalable deployment.

Abstract

Most existing bundle generation approaches fall short in generating fixed-size bundles. Furthermore, they often neglect the underlying user intents reflected by the bundles in the generation process, resulting in less intelligible bundles. This paper addresses these limitations through the exploration of two interrelated tasks, i.e., personalized bundle generation and the underlying intent inference, based on different user sessions. Inspired by the reasoning capabilities of large language models (LLMs), we propose an adaptive in-context learning paradigm, which allows LLMs to draw tailored lessons from related sessions as demonstrations, enhancing the performance on target sessions. Specifically, we first employ retrieval augmented generation to identify nearest neighbor sessions, and then carefully design prompts to guide LLMs in executing both tasks on these neighbor sessions. To tackle reliability and hallucination challenges, we further introduce (1) a self-correction strategy promoting mutual improvements of the two tasks without supervision signals and (2) an auto-feedback mechanism for adaptive supervision based on the distinct mistakes made by LLMs on different neighbor sessions. Thereby, the target session can gain customized lessons for improved performance by observing the demonstrations of its neighbor sessions. Experiments on three real-world datasets demonstrate the effectiveness of our proposed method.
Paper Structure (29 sections, 7 figures, 7 tables)

This paper contains 29 sections, 7 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Recommendation with Prebuilt Bundles
  4. Recommendation with Bundle Generation
  5. Intent-Aware Session Recommendation
  6. LLMs for Recommendation
  7. Prompting Methods for LLMs
  8. The Proposed Methodology
  9. Neighbor Session Retrieval (NSR)
  10. Adaptive Demonstration Generation (ADG)
  11. Mutual Self-Correction
  12. Adaptive Auto-Feedback
  13. Rules Summarization
  14. Demonstration Guided Inference (DGI)
  15. Complexity Discussion
  16. ...and 14 more sections

Figures (7)

  • Figure 1: Example bundles for (1) a camera and its accessories; and (2) mystery, thriller, and historical fiction.
  • Figure 2: The overall framework of AICL. We take one test session and its top-1 neighbor session as an example for illustration.
  • Figure 3: Human evaluation on inferred intents. The bar and horizontal line are mean and standard deviation values, respectively.
  • Figure 4: The impact of $T_b$ on all neighbor sessions.
  • Figure 5: The generated bundles and intents on Electronic.
  • ...and 2 more figures