Implicit feedback -- the main data source for training Recommender Systems (RSs) -- is inherently noisy and has been shown to negatively affect recommendation effectiveness. Denoising has been proposed as a method for removing noisy implicit feedback and improving recommendations. Prior work has focused on in-training denoising, however this requires additional data, changes to the model architecture and training procedure or fine-tuning, all of which can be costly and data hungry. In this work, we focus on post-training denoising. Different from in-training denoising, post-training denoising does not involve changing the architecture of the model nor its training procedure, and does not require additional data. Specifically, we present a method for post-training denoising user profiles using Large Language Models (LLMs) for Collaborative Filtering (CF) recommendations. Our approach prompts LLMs with (i) a user profile (user interactions), (ii) a candidate item, and (iii) its rank as given by the CF recommender, and asks the LLM to remove items from the user profile to improve the rank of the candidate item. Experiments with a state-of-the-art CF recommender and 4 open and closed source LLMs in 3 datasets show that our denoising yields improvements up to 13% in effectiveness over the original user profiles. Our code is available at https://github.com/edervishaj/denoising-user-profiles-LLM.

翻译：隐式反馈——训练推荐系统（RSs）的主要数据来源——本质上是含噪声的，并且已被证明会对推荐效果产生负面影响。去噪作为一种去除噪声隐式反馈以提升推荐效果的方法被提出。先前的研究主要集中于训练中（in-training）去噪，但这需要额外数据、改变模型架构与训练流程或进行微调，这些操作通常成本高昂且数据需求量大。本文聚焦于后训练（post-training）去噪。与训练中去噪不同，后训练去噪既不涉及改变模型架构或其训练流程，也无需额外数据。具体而言，我们提出了一种利用大语言模型（LLMs）对协同过滤（CF）推荐中的用户画像进行后训练去噪的方法。我们的方法向LLMs输入：（i）用户画像（用户交互记录）、（ii）候选物品，以及（iii）CF推荐器给出的该物品排名，并请求LLM从用户画像中移除部分物品以提升候选物品的排名。在一个前沿CF推荐器与4个开源及闭源LLMs上，基于3个数据集的实验表明，我们的去噪方法相比原始用户画像，在推荐效果上最高可提升13%。代码发布于 https://github.com/edervishaj/denoising-user-profiles-LLM。