Searcharxiv⌕ Search

arXiv subjects

Lei Sang

Publications and source records attributed to Lei Sang.

At least 19 recordsLinked to original sources
arXiv2026

GraphLoRA: Structure-Aware Low-Rank Adaptation for Large Language Model Recommendation

Large Language Models (LLMs) have shown strong potential for recommendation (LLMRec) due to their powerful reasoning and generalization abilities. However, effectively aligning the textual semantics modeled by LLMs with the collaborative signals remains a key challenge. Existing methods either translate collaborative information into textual prompts or inject pre-trained embeddings into the LLM, both of which treat structural information as static input and fail to capture high-order relational dependencies. To bridge this gap, we propose GraphLoRA, a novel framework that generalizes low-rank adaptation from independent to structure-aware propagation. GraphLoRA embeds a trainable graph message-passing network within the low-rank adaptation pathway, enabling structural signals to propagate through the parameter space. This design allows collaborative topology to explicitly guide parameter updates, fostering deep integration between graph-structured and textual semantic information. Extensive experiments on multiple benchmarks demonstrate that GraphLoRA not only outperforms state-of-the-art LLM-based recommendation methods but also achieves superior generalization, effectively balancing structural reasoning capability with computational efficiency. Code is available at \href{https://github.com/wgj15965/GraphLoRA}{https://github.com/wgj15965/GraphLoRA}.

cs.CL↗
arXiv2026

DIAURec: Dual-Intent Space Representation Optimization for Recommendation

General recommender systems deliver personalized services by learning user and item representations, with the central challenge being how to capture latent user preferences. However, representations derived from sparse interactions often fail to comprehensively characterize user behaviors, thereby limiting recommendation effectiveness. Recent studies attempt to enhance user representations through sophisticated modeling strategies ($e.g.,$ intent or language modeling). Nevertheless, most works primarily concentrate on model interpretability instead of representation optimization. This imbalance has led to limited progress, as representation optimization is crucial for recommendation quality by promoting the affinity between users and their interacted items in the feature space, yet remains largely overlooked. To overcome these limitations, we propose DIAURec, a novel representation learning framework that unifies intent and language modeling for recommendation. DIAURec reconstructs representations based on the prototype and distribution intent spaces formed by collaborative and language signals. Furthermore, we design a comprehensive representation optimization strategy. Specifically, we adopts alignment and uniformity as the primary optimization objectives, and incorporates both coarse- and fine-grained matching to achieve effective alignment across different spaces, thereby enhancing representational consistency. Additionally, we further introduce intra-space and interaction regularization to enhance model robustness and prevent representation collapse in reconstructed space representation. Experiments on three public datasets against fifteen baseline methods show that DIAURec consistently outperforms state-of-the-art baselines, fully validating its effectiveness and superiority.

cs.IR↗
arXiv2026

From Clues to Generation: Language-Guided Conditional Diffusion for Cross-Domain Recommendation

Cross-domain Recommendation (CDR) exploits multi-domain correlations to alleviate data sparsity. As a core task within this field, inter-domain recommendation focuses on predicting preferences for users who interact in a source domain but lack behavioral records in a target domain. Existing approaches predominantly rely on overlapping users as anchors for knowledge transfer. In real-world scenarios, overlapping users are often scarce, leaving the vast majority of users with only single-domain interactions. For these users, the absence of explicit alignment signals makes fine-grained preference transfer intrinsically difficult. To address this challenge, this paper proposes Language-Guided Conditional Diffusion for CDR (LGCD), a novel framework that integrates Large Language Models (LLMs) and diffusion models for inter-domain sequential recommendation. Specifically, we leverage LLM reasoning to bridge the domain gap by inferring potential target preferences for single-domain users and mapping them to real items, thereby constructing pseudo-overlapping data. We distinguish between real and pseudo-interaction pathways and introduce additional supervision constraints to mitigate the semantic noise brought by pseudo-interaction. Furthermore, we design a conditional diffusion architecture to precisely guide the generation of target user representations based on source-domain patterns. Extensive experiments demonstrate that LGCD significantly outperforms state-of-the-art methods in inter-domain recommendation tasks.

cs.IR↗
arXiv2026

TalkLoRA: Communication-Aware Mixture of Low-Rank Adaptation for Large Language Models

Low-Rank Adaptation (LoRA) enables parameter-efficient fine-tuning of Large Language Models (LLMs), and recent Mixture-of-Experts (MoE) extensions further enhance flexibility by dynamically combining multiple LoRA experts. However, existing MoE-augmented LoRA methods assume that experts operate independently, often leading to unstable routing, expert dominance. In this paper, we propose \textbf{TalkLoRA}, a communication-aware MoELoRA framework that relaxes this independence assumption by introducing expert-level communication prior to routing. TalkLoRA equips low-rank experts with a lightweight Talking Module that enables controlled information exchange across expert subspaces, producing a more robust global signal for routing. Theoretically, we show that expert communication smooths routing dynamics by mitigating perturbation amplification while strictly generalizing existing MoELoRA architectures. Empirically, TalkLoRA consistently outperforms vanilla LoRA and MoELoRA across diverse language understanding and generation tasks, achieving higher parameter efficiency and more balanced expert routing under comparable parameter budgets. These results highlight structured expert communication as a principled and effective enhancement for MoE-based parameter-efficient adaptation. Code is available at https://github.com/why0129/TalkLoRA.

cs.LG↗
arXiv2025

Robustness of Prompting: Enhancing Robustness of Large Language Models Against Prompting Attacks

Large Language Models (LLMs) have demonstrated remarkable performance across various tasks by effectively utilizing a prompting strategy. However, they are highly sensitive to input perturbations, such as typographical errors or slight character order errors, which can significantly impair their performance. Despite advances in prompting techniques such as Chain-of-Thought and automatic prompt generation, developing a prompting strategy that explicitly mitigates the negative impact of such perturbations remains an open challenge. To bridge this gap, we propose Robustness of Prompting (RoP), a novel prompting strategy aimed at enhancing the robustness of LLMs. RoP consists of two stages: Error Correction and Guidance. In the Error Correction stage, RoP applies diverse perturbation methods to generate adversarial examples, which are used to generate prompts that correct input errors automatically. In the Guidance stage, RoP generates an optimal guidance prompt based on the corrected input, guiding the model to generate more robust and accurate inferences. Through comprehensive experiments spanning arithmetic, commonsense, and logical reasoning tasks, we demonstrate that RoP significantly improves LLMs' robustness against adversarial perturbations. Crucially, it preserves model accuracy with only minimal degradation compared to clean input scenarios, thereby establishing RoP as a practical and effective approach for enhancing LLM robustness in real-world applications.

cs.CL↗
arXiv2025

Heterogeneous Graph Masked Contrastive Learning for Robust Recommendation

Heterogeneous graph neural networks (HGNNs) have demonstrated their superiority in exploiting auxiliary information for recommendation tasks. However, graphs constructed using meta-paths in HGNNs are usually too dense and contain a large number of noise edges. The propagation mechanism of HGNNs propagates even small amounts of noise in a graph to distant neighboring nodes, thereby affecting numerous node embeddings. To address this limitation, we introduce a novel model, named Masked Contrastive Learning (MCL), to enhance recommendation robustness to noise. MCL employs a random masking strategy to augment the graph via meta-paths, reducing node sensitivity to specific neighbors and bolstering embedding robustness. Furthermore, MCL employs contrastive cross-view on a Heterogeneous Information Network (HIN) from two perspectives: one-hop neighbors and meta-path neighbors. This approach acquires embeddings capturing both local and high-order structures simultaneously for recommendation. Empirical evaluations on three real-world datasets confirm the superiority of our approach over existing recommendation methods.

cs.IR↗
arXiv2025

Revisiting Feature Interactions from the Perspective of Quadratic Neural Networks for Click-through Rate Prediction

Hadamard Product (HP) has long been a cornerstone in click-through rate (CTR) prediction tasks due to its simplicity, effectiveness, and ability to capture feature interactions without additional parameters. However, the underlying reasons for its effectiveness remain unclear. In this paper, we revisit HP from the perspective of Quadratic Neural Networks (QNN), which leverage quadratic interaction terms to model complex feature relationships. We further reveal QNN's ability to expand the feature space and provide smooth nonlinear approximations without relying on activation functions. Meanwhile, we find that traditional post-activation does not further improve the performance of the QNN. Instead, mid-activation is a more suitable alternative. Through theoretical analysis and empirical evaluation of 25 QNN neuron formats, we identify a good-performing variant and make further enhancements on it. Specifically, we propose the Multi-Head Khatri-Rao Product as a superior alternative to HP and a Self-Ensemble Loss with dynamic ensemble capability within the same network to enhance computational efficiency and performance. Ultimately, we propose a novel neuron format, QNN-alpha, which is tailored for CTR prediction tasks. Experimental results show that QNN-alpha achieves new state-of-the-art performance on six public datasets while maintaining low inference latency, good scalability, and excellent compatibility. The code, running logs, and detailed hyperparameter configurations are available at: https://github.com/salmon1802/QNN.

cs.IR↗
arXiv2025

Quadratic Interest Network for Multimodal Click-Through Rate Prediction

Multimodal click-through rate (CTR) prediction is a key technique in industrial recommender systems. It leverages heterogeneous modalities such as text, images, and behavioral logs to capture high-order feature interactions between users and items, thereby enhancing the system's understanding of user interests and its ability to predict click behavior. The primary challenge in this field lies in effectively utilizing the rich semantic information from multiple modalities while satisfying the low-latency requirements of online inference in real-world applications. To foster progress in this area, the Multimodal CTR Prediction Challenge Track of the WWW 2025 EReL@MIR Workshop formulates the problem into two tasks: (1) Task 1 of Multimodal Item Embedding: this task aims to explore multimodal information extraction and item representation learning methods that enhance recommendation tasks; and (2) Task 2 of Multimodal CTR Prediction: this task aims to explore what multimodal recommendation model can effectively leverage multimodal embedding features and achieve better performance. In this paper, we propose a novel model for Task 2, named Quadratic Interest Network (QIN) for Multimodal CTR Prediction. Specifically, QIN employs adaptive sparse target attention to extract multimodal user behavior features, and leverages Quadratic Neural Networks to capture high-order feature interactions. As a result, QIN achieved an AUC of 0.9798 on the leaderboard and ranked second in the competition. The model code, training logs, hyperparameter configurations, and checkpoints are available at https://github.com/salmon1802/QIN.

cs.IR↗
arXiv2025

Unveiling Contrastive Learning's Capability of Neighborhood Aggregation for Collaborative Filtering

Personalized recommendation is widely used in the web applications, and graph contrastive learning (GCL) has gradually become a dominant approach in recommender systems, primarily due to its ability to extract self-supervised signals from raw interaction data, effectively alleviating the problem of data sparsity. A classic GCL-based method typically uses data augmentation during graph convolution to generates more contrastive views, and performs contrast on these new views to obtain rich self-supervised signals. Despite this paradigm is effective, the reasons behind the performance gains remain a mystery. In this paper, we first reveal via theoretical derivation that the gradient descent process of the CL objective is formally equivalent to graph convolution, which implies that CL objective inherently supports neighborhood aggregation on interaction graphs. We further substantiate this capability through experimental validation and identify common misconceptions in the selection of positive samples in previous methods, which limit the potential of CL objective. Based on this discovery, we propose the Light Contrastive Collaborative Filtering (LightCCF) method, which introduces a novel neighborhood aggregation objective to bring users closer to all interacted items while pushing them away from other positive pairs, thus achieving high-quality neighborhood aggregation with very low time complexity. On three highly sparse public datasets, the proposed method effectively aggregate neighborhood information while preventing graph over-smoothing, demonstrating significant improvements over existing GCL-based counterparts in both training efficiency and recommendation accuracy. Our implementations are publicly accessible.

cs.IR↗
arXiv2025

Intent Representation Learning with Large Language Model for Recommendation

Intent-based recommender systems have garnered significant attention for uncovering latent fine-grained preferences. Intents, as underlying factors of interactions, are crucial for improving recommendation interpretability. Most methods define intents as learnable parameters updated alongside interactions. However, existing frameworks often overlook textual information (e.g., user reviews, item descriptions), which is crucial for alleviating the sparsity of interaction intents. Exploring these multimodal intents, especially the inherent differences in representation spaces, poses two key challenges: i) How to align multimodal intents and effectively mitigate noise issues; ii) How to extract and match latent key intents across modalities. To tackle these challenges, we propose a model-agnostic framework, Intent Representation Learning with Large Language Model (IRLLRec), which leverages large language models (LLMs) to construct multimodal intents and enhance recommendations. Specifically, IRLLRec employs a dual-tower architecture to learn multimodal intent representations. Next, we propose pairwise and translation alignment to eliminate inter-modal differences and enhance robustness against noisy input features. Finally, to better match textual and interaction-based intents, we employ momentum distillation to perform teacher-student learning on fused intent representations. Empirical evaluations on three datasets show that our IRLLRec framework outperforms baselines.Code available at https://github.com/wangyu0627/IRLLRec.

cs.IR↗
arXiv2024

From Collapse to Stability: A Knowledge-Driven Ensemble Framework for Scaling Up Click-Through Rate Prediction Models

Click-through rate (CTR) prediction plays a crucial role in modern recommender systems. While many existing methods utilize ensemble networks to improve CTR model performance, they typically restrict the ensemble to only two or three sub-networks. Whether increasing the number of sub-networks consistently enhances CTR model performance to align with scaling laws remains unclear. In this paper, we investigate larger ensemble networks and find three inherent limitations in commonly used ensemble methods: (1) performance degradation as the number of sub-networks increases; (2) sharp declines and high variance in sub-network performance; and (3) significant discrepancies between sub-network and ensemble predictions. Meanwhile, we analyze the underlying causes of these limitations from the perspective of dimensional collapse: the collapse within sub-networks becomes increasingly severe as the number of sub-networks grows, leading to a lower knowledge abundance. In this paper, we employ knowledge transfer methods, such as Knowledge Distillation (KD) and Deep Mutual Learning (DML), to address the aforementioned limitations. We find that KD enables CTR models to better follow scaling laws, while DML reduces variance among sub-networks and minimizes discrepancies with ensemble predictions. Furthermore, by combining KD and DML, we propose a model-agnostic and hyperparameter-free Knowledge-Driven Ensemble Framework (KDEF) for CTR Prediction.

cs.IR↗
arXiv2024

Feature Interaction Fusion Self-Distillation Network For CTR Prediction

Click-Through Rate (CTR) prediction plays a vital role in recommender systems, online advertising, and search engines. Most of the current approaches model feature interactions through stacked or parallel structures, with some employing knowledge distillation for model compression. However, we observe some limitations with these approaches: (1) In parallel structure models, the explicit and implicit components are executed independently and simultaneously, which leads to insufficient information sharing within the feature set. (2) The introduction of knowledge distillation technology brings about the problems of complex teacher-student framework design and low knowledge transfer efficiency. (3) The dataset and the process of constructing high-order feature interactions contain significant noise, which limits the model's effectiveness. To address these limitations, we propose FSDNet, a CTR prediction framework incorporating a plug-and-play fusion self-distillation module. Specifically, FSDNet forms connections between explicit and implicit feature interactions at each layer, enhancing the sharing of information between different features. The deepest fusion layer is then used as the teacher model, utilizing self-distillation to guide the training of shallow layers. Empirical evaluation across four benchmark datasets validates the framework's efficacy and generalization capabilities. The code is available on https://anonymous.4open.science/r/FSDNet.

cs.IR↗
arXiv2024

Adaptive Fusion Self-supervised Learning for Recommendation

Self-supervised learning (SSL) has recently attracted significant attention in the field of recommender systems. Contrastive learning (CL) stands out as a major SSL paradigm due to its robust ability to generate self-supervised signals. Mainstream graph contrastive learning (GCL)-based methods typically implement CL by creating contrastive views through various data augmentations. Despite these methods are effective, we argue that there still exist several challenges. i) Data augmentation requires additional graph convolution (GCN) or modeling operations, significantly increasing time costs. Moreover, graph augmentation disrupts the intrinsic properties of the user-item graph by randomly removing nodes/edges, while feature augmentation applies noise to all nodes, neglecting their unique characteristics. ii) Existing GCL-based methods use traditional CL objectives to capture self-supervised signals. However, few studies have explored obtaining more beneficial CL objectives from more perspectives and have attempted to fuse the varying self-supervised signals from these CL objectives to enhance recommendation performance. To overcome these challenges, we propose Adaptive Fusion Graph Contrastive Learning (AFGCL) for recommendation. AFGCL exploits structural information naturally produced during graph propagation to construct contrastive representations. Specifically, we introduce an adaptive fusion strategy that estimates the contributions of different propagation depths to the primary recommendation task and adaptively combines their representations. Furthermore, we construct an explicit representation for each observed user--item interaction and propose a fused contrastive objective. Experimental results on three public datasets demonstrate the superior recommendation performance and training efficiency of AFGCL compared with state-of-the-art baselines.

cs.IR↗
arXiv2024

Intent-guided Heterogeneous Graph Contrastive Learning for Recommendation

Contrastive Learning (CL)-based recommender systems have gained prominence in the context of Heterogeneous Graph (HG) due to their capacity to enhance the consistency of representations across different views. However, existing frameworks often neglect the fact that user-item interactions within HG are governed by diverse latent intents (e.g., brand preferences or demographic characteristics of item audiences), which are pivotal in capturing fine-grained relations. The exploration of these underlying intents, particularly through the lens of meta-paths in HGs, presents us with two principal challenges: i) How to integrate CL with intents; ii) How to mitigate noise from meta-path-driven intents. To address these challenges, we propose an innovative framework termed Intent-guided Heterogeneous Graph Contrastive Learning (IHGCL), which designed to enhance CL-based recommendation by capturing the intents contained within meta-paths. Specifically, the IHGCL framework includes: i) a meta-path-based Dual Contrastive Learning (DCL) approach to effectively integrate intents into the recommendation, constructing intent-intent contrast and intent-interaction contrast; ii) a Bottlenecked AutoEncoder (BAE) that combines mask propagation with the information bottleneck principle to significantly reduce noise perturbations introduced by meta-paths. Empirical evaluations conducted across six distinct datasets demonstrate the superior performance of our IHGCL framework relative to conventional baseline methods. Our model implementation is available at https://github.com/wangyu0627/IHGCL.

cs.IR↗
arXiv2024

FCN: Fusing Exponential and Linear Cross Network for Click-Through Rate Prediction

As an important modeling paradigm in click-through rate (CTR) prediction, the Deep & Cross Network (DCN) and its derivative models have gained widespread recognition primarily due to their success in a trade-off between computational cost and performance. This paradigm employs a cross network to explicitly model feature interactions with linear growth, while leveraging deep neural networks (DNN) to implicitly capture higher-order feature interactions. However, these models still face several key limitations: (1) The performance of existing explicit feature interaction methods lags behind that of implicit DNN, resulting in overall model performance being dominated by the DNN; (2) While these models claim to capture high-order feature interactions, they often overlook potential noise within these interactions; (3) The learning process for different interaction network branches lacks appropriate supervision signals; and (4) The high-order feature interactions captured by these models are often implicit and non-interpretable due to their reliance on DNN. To address the identified limitations, this paper proposes a novel model, called Fusing Cross Network (FCN), along with two sub-networks: Linear Cross Network (LCN) and Exponential Cross Network (ECN). FCN explicitly captures feature interactions with both linear and exponential growth, eliminating the need to rely on implicit DNN. Moreover, we introduce the Self-Mask operation to filter noise layer by layer and reduce the number of parameters in the cross network by half. To effectively train these two cross networks, we propose a simple yet effective loss function called Tri-BCE, which provides tailored supervision signals for each network. We evaluate the effectiveness, efficiency, and interpretability of FCN on six benchmark datasets. Furthermore, by integrating LCN and ECN, FCN achieves a new state-of-the-art performance.

cs.IR↗
arXiv2024

Exploring the Individuality and Collectivity of Intents behind Interactions for Graph Collaborative Filtering

Intent modeling has attracted widespread attention in recommender systems. As the core motivation behind user selection of items, intent is crucial for elucidating recommendation results. The current mainstream modeling method is to abstract the intent into unknowable but learnable shared or non-shared parameters. Despite considerable progress, we argue that it still confronts the following challenges: firstly, these methods only capture the coarse-grained aspects of intent, ignoring the fact that user-item interactions will be affected by collective and individual factors (e.g., a user may choose a movie because of its high box office or because of his own unique preferences); secondly, modeling believable intent is severely hampered by implicit feedback, which is incredibly sparse and devoid of true semantics. To address these challenges, we propose a novel recommendation framework designated as Bilateral Intent-guided Graph Collaborative Filtering (BIGCF). Specifically, we take a closer look at user-item interactions from a causal perspective and put forth the concepts of individual intent-which signifies private preferences-and collective intent-which denotes overall awareness. To counter the sparsity of implicit feedback, the feature distributions of users and items are encoded via a Gaussian-based graph generation strategy, and we implement the recommendation process through bilateral intent-guided graph reconstruction re-sampling. Finally, we propose graph contrastive regularization for both interaction and intent spaces to uniformize users, items, intents, and interactions in a self-supervised and non-augmented paradigm. Experimental results on three real-world datasets demonstrate the effectiveness of BIGCF compared with existing solutions.

cs.IR↗
arXiv2024

Dual-domain Collaborative Denoising for Social Recommendation

Social recommendation leverages social network to complement user-item interaction data for recommendation task, aiming to mitigate the data sparsity issue in recommender systems. However, existing social recommendation methods encounter the following challenge: both social network and interaction data contain substaintial noise, and the propagation of such noise through Graph Neural Networks (GNNs) not only fails to enhance recommendation performance but may also interfere with the model's normal training. Despite the importance of denoising for social network and interaction data, only a limited number of studies have considered the denoising for social network and all of them overlook that for interaction data, hindering the denoising effect and recommendation performance. Based on this, we propose a novel model called Dual-domain Collaborative Denoising for Social Recommendation ($\textbf{DCDSR}$). DCDSR comprises two primary modules: the structure-level collaborative denoising module and the embedding-space collaborative denoising module. In the structure-level collaborative denoising module, information from interaction domain is first employed to guide social network denoising. Subsequently, the denoised social network is used to supervise the denoising for interaction data. The embedding-space collaborative denoising module devotes to resisting the noise cross-domain diffusion problem through contrastive learning with dual-domain embedding collaborative perturbation. Additionally, a novel contrastive learning strategy, named Anchor-InfoNCE, is introduced to better harness the denoising capability of contrastive learning. Evaluating our model on three real-world datasets verifies that DCDSR has a considerable denoising effect, thus outperforms the state-of-the-art social recommendation methods.

cs.IR↗
arXiv2024

TF4CTR: Twin Focus Framework for CTR Prediction via Adaptive Sample Differentiation

Effective feature interaction modeling is critical for enhancing the accuracy of click-through rate (CTR) prediction in industrial recommender systems. Most of the current deep CTR models resort to building complex network architectures to better capture intricate feature interactions or user behaviors. However, we identify two limitations in these models: (1) the samples given to the model are undifferentiated, which may lead the model to learn a larger number of easy samples in a single-minded manner while ignoring a smaller number of hard samples, thus reducing the model's generalization ability; (2) differentiated feature interaction encoders are designed to capture different interactions information but receive consistent supervision signals, thereby limiting the effectiveness of the encoder. To bridge the identified gaps, this paper introduces a novel CTR prediction framework by integrating the plug-and-play Twin Focus (TF) Loss, Sample Selection Embedding Module (SSEM), and Dynamic Fusion Module (DFM), named the Twin Focus Framework for CTR (TF4CTR). Specifically, the framework employs the SSEM at the bottom of the model to differentiate between samples, thereby assigning a more suitable encoder for each sample. Meanwhile, the TF Loss provides tailored supervision signals to both simple and complex encoders. Moreover, the DFM dynamically fuses the feature interaction information captured by the encoders, resulting in more accurate predictions. Experiments on five real-world datasets confirm the effectiveness and compatibility of the framework, demonstrating its capacity to enhance various representative baselines in a model-agnostic manner. To facilitate reproducible research, our open-sourced code and detailed running logs will be made available at: https://github.com/salmon1802/TF4CTR.

cs.IR↗