Do LLMs Implicitly Exhibit User Discrimination in Recommendation? An Empirical Study

Recently, Large Language Models (LLMs) have enhanced user interaction, enabling seamless information retrieval and recommendations. However, concerns emerge as these LLMs have shown tendencies to display discrimination related to users' sensitive characteristics (such as gender), leading to explicit user unfairness. Furthermore, our analysis uncovers a more discreet variant of bias in LLMs, defined as implicit user unfairness, wherein these models demonstrate discriminatory recommendation behaviors based solely on non-sensitive user details, like usernames or email addresses. This subtle form of unfairness, while more pervasive, poses a significant threat to the ethical integrity and rights of minority user groups. To comprehensively explore implicit user unfairness, our analysis unfolds in three key steps: (1) We uncover the reasons for this implicit user unfairness: LLMs can infer users' sensitive attributes from non-sensitive attributes (e.g. user names) due to their extensive world knowledge. (2) Our findings expose that the magnitude of implicit user unfairness within LLMs surpasses the level of explicit user unfairness observed in traditional recommender models, signifying a more alarming issue of unfairness, i.e. some non-sensitive features of users like names may result in more serious discrimination phenomena. (3) We analyze the long-term effect of implicit user unfairness, identifying that it will reinforce information bubbles at an accelerated rate compared to traditional RS. We emphasize the need to identify and mitigate implicit user unfairness, aiming to avert the potential human-LLMs recommendation systems deterioration.

Attack Prompt Generation for Red Teaming and Defending Large Language Models

Large language models (LLMs) are susceptible to red teaming attacks, which can induce LLMs to generate harmful content. Previous research constructs attack prompts via manual or automatic methods, which have their own limitations on construction cost and quality. To address these issues, we propose an integrated approach that combines manual and automatic methods to economically generate high-quality attack prompts. Specifically, considering the impressive capabilities of newly emerged LLMs, we propose an attack framework to instruct LLMs to mimic human-generated prompts through in-context learning. Furthermore, we propose a defense framework that fine-tunes victim LLMs through iterative interactions with the attack framework to enhance their safety against red teaming attacks. Extensive experiments on different LLMs validate the effectiveness of our proposed attack and defense frameworks. Additionally, we release a series of attack prompts datasets named SAP with varying sizes, facilitating the safety evaluation and enhancement of more LLMs. Our code and dataset is available on https://github.com/Aatrox103/SAP .

A Multi-facet Paradigm to Bridge Large Language Model and Recommendation

Large Language Models (LLMs) have garnered considerable attention in recommender systems. To achieve LLM-based recommendation, item indexing and generation grounding are two essential steps, bridging between recommendation items and natural language. Item indexing assigns a unique identifier to represent each item in natural language, and generation grounding grounds the generated token sequences to in-corpus items. However, previous works suffer from inherent limitations in the two steps. For item indexing, existing ID-based identifiers (e.g., numeric IDs) and description-based identifiers (e.g., titles) often compromise semantic richness or uniqueness. Moreover, generation grounding might inadvertently produce out-of-corpus identifiers. Worse still, autoregressive generation heavily relies on the initial token's quality. To combat these issues, we propose a novel multi-facet paradigm, namely TransRec, to bridge the LLMs to recommendation. Specifically, TransRec employs multi-facet identifiers that incorporate ID, title, and attribute, achieving both distinctiveness and semantics. Additionally, we introduce a specialized data structure for TransRec to guarantee the in-corpus identifier generation and adopt substring indexing to encourage LLMs to generate from any position. We implement TransRec on two backbone LLMs, i.e., BART-large and LLaMA-7B. Empirical results on three real-world datasets under diverse settings (e.g., full training and few-shot training with warm- and cold-start testings) attest to the superiority of TransRec.

RecAD: Towards A Unified Library for Recommender Attack and Defense

In recent years, recommender systems have become a ubiquitous part of our daily lives, while they suffer from a high risk of being attacked due to the growing commercial and social values. Despite significant research progress in recommender attack and defense, there is a lack of a widely-recognized benchmarking standard in the field, leading to unfair performance comparison and limited credibility of experiments. To address this, we propose RecAD, a unified library aiming at establishing an open benchmark for recommender attack and defense. RecAD takes an initial step to set up a unified benchmarking pipeline for reproducible research by integrating diverse datasets, standard source codes, hyper-parameter settings, running logs, attack knowledge, attack budget, and evaluation results. The benchmark is designed to be comprehensive and sustainable, covering both attack, defense, and evaluation tasks, enabling more researchers to easily follow and contribute to this promising field. RecAD will drive more solid and reproducible research on recommender systems attack and defense, reduce the redundant efforts of researchers, and ultimately increase the credibility and practical value of recommender attack and defense. The project is released at https://github.com/gusye1234/recad.

A Bi-Step Grounding Paradigm for Large Language Models in Recommendation Systems

As the focus on Large Language Models (LLMs) in the field of recommendation intensifies, the optimization of LLMs for recommendation purposes (referred to as LLM4Rec) assumes a crucial role in augmenting their effectiveness in providing recommendations. However, existing approaches for LLM4Rec often assess performance using restricted sets of candidates, which may not accurately reflect the models' overall ranking capabilities. In this paper, our objective is to investigate the comprehensive ranking capacity of LLMs and propose a two-step grounding framework known as BIGRec (Bi-step Grounding Paradigm for Recommendation). It initially grounds LLMs to the recommendation space by fine-tuning them to generate meaningful tokens for items and subsequently identifies appropriate actual items that correspond to the generated tokens. By conducting extensive experiments on two datasets, we substantiate the superior performance, capacity for handling few-shot scenarios, and versatility across multiple domains exhibited by BIGRec. Furthermore, we observe that the marginal benefits derived from increasing the quantity of training samples are modest for BIGRec, implying that LLMs possess the limited capability to assimilate statistical information, such as popularity and collaborative filtering, due to their robust semantic priors. These findings also underline the efficacy of integrating diverse statistical information into the LLM4Rec framework, thereby pointing towards a potential avenue for future research. Our code and data are available at https://github.com/SAI990323/Grounding4Rec.

General Debiasing for Multimodal Sentiment Analysis

Existing work on Multimodal Sentiment Analysis (MSA) utilizes multimodal information for prediction yet unavoidably suffers from fitting the spurious correlations between multimodal features and sentiment labels. For example, if most videos with a blue background have positive labels in a dataset, the model will rely on such correlations for prediction, while "blue background" is not a sentiment-related feature. To address this problem, we define a general debiasing MSA task, which aims to enhance the Out-Of-Distribution (OOD) generalization ability of MSA models by reducing their reliance on spurious correlations. To this end, we propose a general debiasing framework based on Inverse Probability Weighting (IPW), which adaptively assigns small weights to the samples with larger bias (i.e., the severer spurious correlations). The key to this debiasing framework is to estimate the bias of each sample, which is achieved by two steps: 1) disentangling the robust features and biased features in each modality, and 2) utilizing the biased features to estimate the bias. Finally, we employ IPW to reduce the effects of large-biased samples, facilitating robust feature learning for sentiment prediction. To examine the model's generalization ability, we keep the original testing sets on two benchmarks and additionally construct multiple unimodal and multimodal OOD testing sets. The empirical results demonstrate the superior generalization ability of our proposed framework. We have released the code and data to facilitate the reproduction https://github.com/Teng-Sun/GEAR.

Robust Instruction Optimization for Large Language Models with Distribution Shifts

Large Language Models have demonstrated significant ability in accomplishing a wide range of Natural Language Processing (NLP) tasks. However, their performance is highly sensitive to the even minor changes in the phrasing of the task instructions, leading to a line of research in automatic instruction optimization towards better performance for NLP tasks. Unfortunately, existing methods for instruction optimization fail to consider the distribution shift between the seen training data and the unseen test data, where testing on unseen group of data with a different distribution could potentially lead to performance drop. In this paper, we take an initial step of investigating the problem of LLM instruction optimization across data groups with distribution shifts. We find that the optimal instructions do encounter performance drops on LLM under certain distribution shifts. To this end, we propose a framework to derive more robust optimal instructions that improve the performance on the unseen data group without large sacrifice on the seen data group. Experimental results demonstrate the effectiveness of our proposed framework.

Is ChatGPT Fair for Recommendation? Evaluating Fairness in Large Language Model Recommendation

The remarkable achievements of Large Language Models (LLMs) have led to the emergence of a novel recommendation paradigm -- Recommendation via LLM (RecLLM). Nevertheless, it is important to note that LLMs may contain social prejudices, and therefore, the fairness of recommendations made by RecLLM requires further investigation. To avoid the potential risks of RecLLM, it is imperative to evaluate the fairness of RecLLM with respect to various sensitive attributes on the user side. Due to the differences between the RecLLM paradigm and the traditional recommendation paradigm, it is problematic to directly use the fairness benchmark of traditional recommendation. To address the dilemma, we propose a novel benchmark called Fairness of Recommendation via LLM (FaiRLLM). This benchmark comprises carefully crafted metrics and a dataset that accounts for eight sensitive attributes1 in two recommendation scenarios: music and movies. By utilizing our FaiRLLM benchmark, we conducted an evaluation of ChatGPT and discovered that it still exhibits unfairness to some sensitive attributes when generating recommendations. Our code and dataset can be found at https://github.com/jizhi-zhang/FaiRLLM.

STAR-RIS Aided Covert Communication

This paper investigates the multi-antenna covert communications assisted by a simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS). In particular, to shelter the existence of communications between transmitter and receiver from a warden, a friendly full-duplex receiver with two antennas is leveraged to make contributions to confuse the warden. Considering the worst case, the closed-form expression of the minimum detection error probability (DEP) at the warden is derived and utilized as a covert constraint. Then, we formulate an optimization problem maximizing the covert rate of the system under the covertness constraint and quality of service (QoS) constraint with communication outage analysis. To jointly design the active and passive beamforming of the transmitter and STAR-RIS, an iterative algorithm based on globally convergent version of method of moving asymptotes (GCMMA) is proposed to effectively solve the non-convex optimization problem. Simulation results show that the proposed STAR-RIS-assisted scheme highly outperforms the case with conventional RIS.