Understanding the Role of Cross-Entropy Loss in Fairly Evaluating Large Language Model-based Recommendation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xu, Cong, Zhu, Zhangchi, Wang, Jun, Wang, Jianyong, Zhang, Wei
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910340255580160
author Xu, Cong
Zhu, Zhangchi
Wang, Jun
Wang, Jianyong
Zhang, Wei
author_facet Xu, Cong
Zhu, Zhangchi
Wang, Jun
Wang, Jianyong
Zhang, Wei
contents Large language models (LLMs) have gained much attention in the recommendation community; some studies have observed that LLMs, fine-tuned by the cross-entropy loss with a full softmax, could achieve state-of-the-art performance already. However, these claims are drawn from unobjective and unfair comparisons. In view of the substantial quantity of items in reality, conventional recommenders typically adopt a pointwise/pairwise loss function instead for training. This substitute however causes severe performance degradation, leading to under-estimation of conventional methods and over-confidence in the ranking capability of LLMs. In this work, we theoretically justify the superiority of cross-entropy, and showcase that it can be adequately replaced by some elementary approximations with certain necessary modifications. The remarkable results across three public datasets corroborate that even in a practical sense, existing LLM-based methods are not as effective as claimed for next-item recommendation. We hope that these theoretical understandings in conjunction with the empirical results will facilitate an objective evaluation of LLM-based recommendation in the future.
format Preprint
id arxiv_https___arxiv_org_abs_2402_06216
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Understanding the Role of Cross-Entropy Loss in Fairly Evaluating Large Language Model-based Recommendation
Xu, Cong
Zhu, Zhangchi
Wang, Jun
Wang, Jianyong
Zhang, Wei
Information Retrieval
Large language models (LLMs) have gained much attention in the recommendation community; some studies have observed that LLMs, fine-tuned by the cross-entropy loss with a full softmax, could achieve state-of-the-art performance already. However, these claims are drawn from unobjective and unfair comparisons. In view of the substantial quantity of items in reality, conventional recommenders typically adopt a pointwise/pairwise loss function instead for training. This substitute however causes severe performance degradation, leading to under-estimation of conventional methods and over-confidence in the ranking capability of LLMs. In this work, we theoretically justify the superiority of cross-entropy, and showcase that it can be adequately replaced by some elementary approximations with certain necessary modifications. The remarkable results across three public datasets corroborate that even in a practical sense, existing LLM-based methods are not as effective as claimed for next-item recommendation. We hope that these theoretical understandings in conjunction with the empirical results will facilitate an objective evaluation of LLM-based recommendation in the future.
title Understanding the Role of Cross-Entropy Loss in Fairly Evaluating Large Language Model-based Recommendation
topic Information Retrieval
url https://arxiv.org/abs/2402.06216