The powerful text understanding and generation capabilities of large language models (LLMs) have brought new vitality to general recommendation with implicit feedback. One possible strategy involves generating a unique user (or item) profile from historical interaction data, which is then mapped to a semantic representation in the language space. However, a single-instance profile may be insufficient to comprehensively capture the complex intentions behind a user's interacted items. Moreover, due to the inherent instability of LLMs, a biased or misinterpreted profile could even undermine the original recommendation performance. Consequently, an intuitive solution is to generate multiple profiles for each user (or item), each reflecting a distinct aspect of their characteristics. In light of this, we propose a unified recommendation framework with multi-faceted profile extrapolation (ProEx) in this paper. By leveraging chain-of-thought reasoning, we construct multiple distinct profiles for each user and item. These new profiles are subsequently mapped into semantic vectors, extrapolating from the position of the original profile to explore a broader region of the language space. Subsequently, we introduce the concept of environments, where each environment represents a possible linear combination of all profiles. The differences across environments are minimized to reveal the inherent invariance of user preferences. We apply ProEx to three discriminative methods and three generative methods, and conduct extensive experiments on three datasets. The experimental results demonstrate that ProEx significantly enhances the performance of these base recommendation models.

The rapid advancement of large language models (LLMs) has opened new possibilities for understanding users in search and recommendation. While traditional behavior-based or feature-driven user models rely primarily on explicit interactions or handcrafted representations, LLMs introduce a fundamentally different paradigm: LLM-powered user profiling, where user preferences, intents, and contextual attributes can be extracted, summarized, or reasoned about directly through natural language. This shift unlocks powerful new paths to achieve personalization but also raises pressing questions related to modeling fidelity, temporal dynamics, evaluation methodology, privacy, and responsible deployment. The LLM-UP workshop aims to bring together researchers and practitioners to systematize emerging progress in LLM-powered user profiling, identify open challenges, and explore opportunities for integrating such techniques into search and recommendation pipelines. The LLM-UP workshop adopts an interactive structure featuring lightning talks, panel discussions, and paper presentations to foster active engagement, cross-disciplinary dialogue, and community-driven agenda setting for this rapidly evolving field.

The remarkable text understanding and generation capabilities of large language models (LLMs) have revitalized the field of general recommendation based on implicit user feedback. Rather than deploying LLMs directly as recommendation models, a more flexible paradigm leverages their ability to interpret users' historical interactions and semantic contexts to extract structured profiles that characterize user preferences. These profiles can be further transformed into actionable high-dimensional representations, serving as powerful signals to augment and strengthen recommendation models. However, the mechanism by which such profiles enhance recommendation performance within the feature space remains insufficiently understood. Moreover, existing studies predominantly rely on nonlinear alignment and fusion strategies to incorporate these profiles, which often lead to semantic loss and fail to fully exploit their potential. To address these limitations, we revisit profiles from a retrieval perspective and propose a simple yet effective recommendation framework built upon distribution shaping (ProMax) in this paper. We begin by employing dense retrieval to uncover the collaborative relationships between user and item profiles within the feature space. Based on this insight, we introduce a dual distribution-reshaping process, in which the profile distribution acts as a guiding signal to steer the recommendation model toward learning user preferences for unseen items beyond the scope of observed interactions. We apply ProMax to four classic recommendation methods on three public datasets. The results indicate that ProMax substantially improves base model performance and outperforms existing LLM-based recommendation approaches.

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.

Graph-based collaborative filtering techniques have emerged as a promising recommendation approach by modeling user-item interaction as graphs. Recently, contrastive learning has been employed in graph collaborative through data augmentation, which can effectively offer data efficiency and reduce labeling costs. Nonetheless, most existing contrastive learning approaches overlook the heterogeneous auxiliary information pertaining to users and items, such as user social relationships and item categories, which are crucial to alleviate the data sparsity issue. In this paper, we propose a novel contrastive learning method, referred to as Heterogeneous Adaptive Preference Learning for Recommendation (HAPLRec), which explicitly incorporates fine-grained preference information from both users and items. Specifically, we construct user relationship graphs and item relationship graphs based on specific meta-paths in a heterogeneous graph. Subsequently, we conduct data augmentation on these graphs individually to obtain auxiliary contrastive tasks. Moreover, we introduce an optimization algorithm that leverages the gradient similarity between the main task and the auxiliary tasks, dynamically adjusting the weight assigned to each task to expedite achieving superior performance within a shorter time frame. The effectiveness of the proposed model is demonstrated through extensive experiments conducted on three publicly available datasets.

Generative recommendation aims to learn the underlying generative process over the entire item set to produce recommendations for users. Although it leverages non-linear probabilistic models to surpass the limited modeling capacity of linear factor models, it is often constrained by a trade-off between representation ability and tractability. With the rise of a new generation of generative methods based on pre-trained language models (LMs), incorporating LMs into general recommendation with implicit feedback has gained considerable attention. However, adapting them to generative recommendation remains challenging. The core reason lies in the mismatch between the input-output formats and semantics of generative models and LMs, making it challenging to achieve optimal alignment in the feature space. This work addresses this issue by proposing a model-agnostic generative recommendation framework called DMRec, which introduces a probabilistic meta-network to bridge the outputs of LMs with user interactions, thereby enabling an equivalent probabilistic modeling process. Subsequently, we design three cross-space distribution matching processes aimed at maximizing shared information while preserving the unique semantics of each space and filtering out irrelevant information. We apply DMRec to three different types of generative recommendation methods and conduct extensive experiments on three public datasets. The experimental results demonstrate that DMRec can effectively enhance the recommendation performance of these generative models, and it shows significant advantages over mainstream LM-enhanced recommendation methods.

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 modelagnostic 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 1 .

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.

The core of the general recommender systems lies in learning high-quality embedding representations of users and items to investigate their positional relations in the feature space. Unfortunately, data sparsity caused by difficult-to-access interaction data severely limits the effectiveness of recommender systems. Faced with such a dilemma, various types of self-supervised learning methods have been introduced into recommender systems in an attempt to alleviate the data sparsity through distribution modeling or data augmentation. However, most data augmentation relies on elaborate manual design, which is not only not universal, but the bloated and redundant augmentation process may significantly slow down model training progress. To tackle these limitations, we propose a novel Dual Mixing-based Recommendation Framework (MixRec) to empower data augmentation as we wish. Specifically, we propose individual mixing and collective mixing, respectively. The former aims to provide a new positive sample that is unique to the target (user or item) and to make the pair-wise recommendation loss benefit from it, while the latter aims to portray a new sample that contains group properties in a batch. The two mentioned mixing mechanisms allow for data augmentation with only one parameter that does not need to be set multiple times and can be done in linear time complexity. Besides, we propose the dual-mixing contrastive learning to maximize the utilization of these new-constructed samples to enhance the consistency between pairs of positive samples. Experimental results on four real-world datasets demonstrate the advantages of MixRec in terms of effectiveness, simplicity, efficiency, and scalability.

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 termedIntent-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.

Click-through rate (CTR) prediction is a crucial task in personalized information retrievals, such as industrial recommender systems, online advertising, and web search. Most existing CTR Prediction models utilize explicit feature interactions to overcome the performance bottleneck of implicit feature interactions. Hence, deep CTR models based on parallel structures (e.g., DCN, FinalMLP, xDeepFM) have been proposed to obtain joint information from different semantic spaces. However, these parallel subcomponents lack effective supervision and communication signals, making it challenging to efficiently capture valuable multi-views feature interaction information in different semantic spaces. To address these issues, we propose a simple yet effective novel CTR model: Contrast-enhanced Through Network (CETN). Drawing inspiration from sociology, CETN leverages the complementary nature of diversity and homogeneity to guide the model in acquiring higher-quality feature interaction information. Specifically, CETN employs product-based feature interactions and the augmentation (perturbation) concept from contrastive learning to segment different semantic spaces, each with distinct activation functions. This improves diversity in the feature interaction information captured by the model. Additionally, we introduce self-supervised signals and through connection within each semantic space to ensure the homogeneity of the captured feature interaction information. The experiments conducted on four real datasets demonstrate that our model consistently outperforms twenty baseline models in terms of AUC and Logloss.

The goal of click-through rate (CTR) prediction in recommender systems is to effectively work with input features. However, existing CTR prediction models face three main issues. First, many models use a basic approach for feature combinations, leading to noise and reduced accuracy. Second, there is no consideration for the varying importance of features in different interaction orders, affecting model performance. Third, current model architectures struggle to capture different interaction signals from various semantic spaces, leading to sub-optimal performance. To address these issues, we propose the Adaptive Graph Interaction Network (AdaGIN) with the Graph Neural Networks-based Feature Interaction Module (GFIM), the Multi-semantic Feature Interaction Module (MFIM), and the Negative Feedback-based Search (NFS) algorithm. GFIM explicitly aggregates information between features and assesses their importance, while MFIM captures information from different semantic spaces. NFS uses negative feedback to optimize model complexity. Experimental results show AdaGIN outperforms existing models on large-scale public benchmark datasets.

The tremendous positive driving effect of Graph Convolutional Network (GCN) and Graph Contrastive Learning (GCL) for recommender systems has become a consensus. GCN encoders are extensively used in recommendation models for capturing high-order connectivities between users and items, whereas GCL accelerates the training of recommendation tasks by adding extra supervision signals from contrastive objectives. However, little attention has been paid on corresponding theories that are truly tailored to recommendation tasks. From the technical perspective, Collaborative Filtering (CF) is seen as an important factor in recommender systems. It is applied to measure user–user, item–item, and user–item similarities rather than to achieve better clustering or node classification results. Besides, heuristic-based data augmentation may not be hold true in the field of recommender systems as it requires additional training costs and introduces noises that will corrupt the interaction graph structure and the semantic information of nodes. To tackle these limitations, we propose a novel Embedding-Less Graph Collaborative Filtering (EGCF) for recommendation, which is tailor-made for the problem mentioned for CF and further simplifies existing solutions. Structurally, it consists of two parts: embedding-less GCN and embedding-less GCL. The former improves user–item affinity by streamlining user-type embeddings and carrying out iterative graph convolution. And the latter utilizes three-type contrastive objectives to directly measure the alignment and the uniformity of users, items, and interaction pairs, respectively, avoiding any type of data augmentation or multi-view construction. Even though EGCF has been extremely streamlined, extensive experimental results on three classical datasets demonstrate the effectiveness of EGCF in terms of recommendation accuracy and training efficiency. The code and used datasets are released at https://github.com/BlueGhostYi/ID-GRec .

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.

As a new recommendation pattern combining collaborative filtering and social network, social recommender system strives to introduce auxiliary user relations to alleviate data sparsity problems. Considering the graph structure characteristics of user historical interactions and social network, there have been emerged several innovative works that utilize Graph Neural Network (GNN) for social recommendation to show impressive performance. However, existing works seem to be restricted to exploiting social network as auxiliary information for main recommendation tasks, with little attention on the social network itself at the fine-grained level. From empirical perspective, the effectiveness of directly applying social network to social recommendation via GNNs may be limited since the social information that can be used for training is actually sparser than user interactions, and most of observable social information is not valid. To resolve this problem, we propose a Dual Variational Graph Reconstruction Learning (DVGRL) framework for social recommendation. It treats user interaction graph and social network as equivalent and aims to learn both variational distributions of user preferences from historical interactions and social connections, which are trained simultaneously and used to guide the reconstruction of historical interaction graph and social network. To effectively exploit the social information gleaned from reconstruction learning for enhancing recommendation, we design two inter-domain fusion mechanisms to achieve knowledge transfer from the perspectives of attention features and prior distributions, respectively. Extensive experiments on four real-world datasets validate the effectiveness of DVGRL for social recommendation tasks.

Graph-based recommender system has attracted widespread attention and produced a series of research results. Because of the powerful high-order connection modeling capabilities of the Graph Neural Network, the performance of these graph-based recommender systems are far superior to those of traditional neural network-based collaborative filtering models. However, from both analytical and empirical perspectives, the apparent performance improvement is accompanied with a significant time overhead, which is noticeable in large-scale graph topologies. More importantly, the intrinsic data-sparsity problem substantially limits the performance of graph-based recommender systems, which compelled us to revisit graph-based recommendation from a novel perspective. In this article, we focus on analyzing the time complexity of graph-based recommender systems to make it more suitable for real large-scale application scenarios. We propose a novel end-to-end graph recommendation model called the Collaborative Variational Graph Auto-Encoder (CVGA), which uses the information propagation and aggregation paradigms to encode user–item collaborative relationships on the user–item interaction bipartite graph. These relationships are utilized to infer the probability distribution of user behavior for parameter estimation rather than learning user or item embeddings. By doing so, we reconstruct the whole user–item interaction graph according to the known probability distribution in a feasible and elegant manner. From the perspective of the graph auto-encoder, we convert the graph recommendation task into a graph generation problem and are able to do it with approximately linear time complexity. Extensive experiments on four real-world benchmark datasets demonstrate that CVGA can be trained at a faster speed while maintaining comparable performance over state-of-the-art baselines for graph-based recommendation tasks. Further analysis shows that CVGA can effectively mitigate the data sparsity problem and performs equally well on large-scale datasets.