Multi-behavior recommendation systems enhance prediction accuracy for target behavior (e.g., purchase) by integrating auxiliary behaviors such as viewing and adding to cart. However, existing models face two fundamental challenges regarding data characteristics: On one hand, they typically treat all auxiliary behaviors equally or apply simple weighting, lacking effective mechanisms to evaluate and filter inherent noise, leading to contaminated feature representations. On the other hand, when fusing sparse target features with abundant auxiliary features, the critical target signal is easily diluted, losing its dominant role in the final prediction. To address this, we propose a Target-Enhanced Gated Transformer model TEGT. Its core innovations include: a behavior-adaptive gating module that filters source-level noise through nonlinear transformations and hard thresholding while generating importance weights; a target-guided dual-modulation attention mechanism utilizes target behavior as queries to retrieve semantically relevant auxiliary signals, then applies secondary modulation by combining gated weights to ensure both noise resistance and target dominance in the fusion process; a lightweight collaborative semantic enhancement module clusters the fused representations and employs cluster-center contrastive learning to explicitly amplify collaborative signals under sparse target behavior. Extensive experiments on three real-world datasets show that TEGT consistently outperforms state-of-the-art baselines, it achieves remarkable improvements of up to 9.73% in Recall@10 and 9.16% in NDCG@10.

Major promotional events such as Black Friday and 618 Shopping Day cause drastic, heterogeneous shifts in user and advertiser behavior, posing persistent challenges for conversion rate (CVR) models trained on daily data. Existing methods often lack the flexibility to capture this periodic variability, resulting in poor modeling of diverse behavioral patterns. To address these challenges, we propose TemporalExpertNet(TEN), a cross-temporal transfer learning framework for industrial-scale CVR prediction during promotion cycles. TEN decomposes the model into a stable representation encoder and a promotion-sensitive expert, enabling reusable temporal knowledge transfer. Specifically, we propose BridgeNet to address the mismatch between historical knowledge and current features through temporal representation alignment. We further introduce TemporalExpertGate (TEG) to perform sample-aware expert fusion, enabling fine-grained prediction adjustment and adaptive knowledge reuse across promotion periods. By using a two-stage training strategy, TEN achieves stable alignment and adaptive expert fusion for robust prediction under shifting promotional distributions. TEN was deployed on a large-scale short-video ads platform during the 618 preheating phase, improving conversion rate by 7.52% and platform RPM by 4.27% with only 0.23% model size and 1.8% latency overhead. It was therefore fully launched to all traffic on 618 Shopping Day, bringing substantial commercial gains.

Existing LLM-driven recommendation systems (RS) suffer from over-reliance on a single pre-trained model, which limits adaptability across diverse scenarios due to differences in large language models' strengths in semantics, knowledge, and reasoning. To address this issue, we propose AMLrec (Adaptive Multi-LLM Recommendation), a dynamic routing-based adaptive multi-LLM collaboration framework that unifies two dominant paradigms—LLM as Recommender and LLM + Recommender—through decision knowledge complementation. For each user or item, a lightweight encoder generates embeddings that are compared with learnable LLM prototypes using cosine similarity to select the most suitable models. In the first paradigm, selected LLMs generate recommendations via structured prompts, and their outputs are aggregated to form the final recommendation list. In the second paradigm, chosen LLMs produce semantic embeddings, which are fused with learnable embeddings after PCA-based dimensionality reduction and aligned using a lightweight adapter to bridge distribution gaps. Notably, AMLrec does not require fine-tuning of the underlying LLMs, significantly reducing computational overhead. Experiments on real-world datasets demonstrate that the proposed approach consistently outperforms single-LLM baselines across all evaluation metrics, validating its effectiveness. The main contributions of this work are threefold: introducing dynamic routing for multi-LLM recommendation system collaboration, proposing a unified architecture that harmonizes both paradigms, and enabling efficient adaptation without LLM fine-tuning. The code is available at https://github.com/Jiale-12138/AMLrec.

Multi-objective evolutionary algorithms (MOEAs) have achieved notable success in recommendation systems (RSs) by meeting diverse user needs. However, existing MOEAs lack effective methods to coordinate the challenges of cold start, low convergence of multiple objectives and lack of explainable recommendation reasons. Therefore, we propose an enhancing multi-objective optimization recommendation algorithm by integrating large language models (named as MORA-LLM). MORA-LLM uses the vast knowledge, reasoning ability and natural language generation (NLG) ability of large language models (LLMs) to compensate for the shortcomings of MOEA-based RSs in semantic understanding. Firstly, an LLM-enhancing prediction score (LEPS) strategy is proposed to alleviate the cold start problem. LEPS obtains the user embedding vector by vast knowledge of LLM and extracts interaction information of similar users to improve the accuracy of prediction scores. Secondly, an LLM-enhancing search (LES) strategy is proposed to improve the convergence of the multi-objective. LES strategy combines the reasoning ability of LLM with the competitive idea of competitive swarm optimization to achieve efficient search and balance multiple objectives. Finally, the prediction scores are further corrected based on the LLM output results and MORA-LLM offers recommendation reasons to help users better understand the recommendation results. Experimental results on real-world datasets demonstrate that MORA-LLM significantly outperforms existing algorithms in terms of recommendation accuracy and convergence.

Multi-party multi-objective optimization, which aims to find a solution set that satisfies multiple decision makers (DMs) as much as possible, has attracted the attention of researchers recently. Although multi-party multi-objective optimization is of great significance in practical applications, most existing works focus on continuous problems while paying little attention to discrete problems. To this end, we propose a multi-party multi-objective evolutionary framework named MP-HCEA for discrete problems, where a multi-party population is used to optimize all objectives of multiple DMs, and multiple single-party populations are used to respectively optimize the objectives of each DM. In MP-HCEA, a dual-phase cooperation mechanism is firstly proposed to guide the population interaction, where the weak cooperation is performed in the early phase to share offspring individuals, while the strong cooperation is performed in the later phase to share parent individuals. This dual-phase cooperation mechanism not only ensures effective information sharing between multiple populations, but also helps them to obtain high-quality solutions. In addition, a novel dual-search mechanism is proposed to guide the evolution of the multi-party population, which further enhances the convergence ability of the algorithm. Finally, we apply MP-HCEA to a real application named multi-stakeholder recommendation as a case study. Experiments on real-world multi-stakeholder recommendation datasets show that the proposed MP-HCEA outperforms several representative baselines.

The proliferation of online publications and interdisciplinary studies has presented researchers with the challenge of sifting through a substantial volume of articles to identify citations that substantiate their research ideas. Consequently, the development of citation recommendation technology has become a pivotal aspect of product promotion and marketing for academic support platforms. Traditionally, citation recommendation has primarily relied either on collaborative signals derived from paper interactions or on content-based similarity—both of which are essential for identifying relevant references. However, existing basic strategies often focus on one of these aspects while neglecting the other, leading to suboptimal performance in capturing the complex factors behind citation behavior. The reason is that the integration of domain characteristics in scholarly fields and the mining of semantic relevance in text information is also crucial for modelling researchers’ preferences. In this work, we present a novel citation recommendation model called SCTRec, that aligns S equential C ollaborative signals from publications’ indexes, that is, IDs, and T ext semantic content for citation Rec ommendation. To address the actual technical challenges encountered, such as shifts in user preferences and semantic gaps between IDs and texts, we have designed a hybrid enhancement mechanism that bridges these semantic gaps, thereby learning more discriminative feature representations. The effectiveness of SCTRec in enhancing citation recommendation performance is substantiated by extensive experimental evaluation on multiple public datasets. The code is available on Anonymous Github at https://github.com/guaiqihen/SCTRec .

Multi-behavior recommender systems improve prediction accuracy of target behaviors (e.g., purchases) by integrating auxiliary behaviors (e.g., page views). However, existing models face two key limitations: (1) Static propagation mechanisms and inflexible dependency modeling fail to capture dynamic changes in user preferences and cascading relationships between behaviors; (2) Sparse target behavior data usually leads to excessive influence of auxiliary signals, which degrades recommendation quality. To address these challenges, we propose the Dynamic Dependency-Aware Collaborative Contrastive Learning Multi-Behavior Recommendation Model, MBDCC. MBDCC has two dedicated modules: (1) Behavioral gating cascade and cross-attention fusion module, which dynamically models cascading dependencies between behaviors through learnable gate control transfer units controlled by behavioral attributes. This replaces static propagation with adaptive feature flow regulation, capturing evolving user preferences. Meanwhile, it uses a target-guided cross-attention mechanism to selectively fuse semantically relevant auxiliary signals using the target behavior as a query, addressing inflexible cross-behavioral dependency modeling; (2) Collaborative semantic enhancement module, it constructs a user similarity measure matrix based on co-occurrence frequency of interaction items in target behavior, and clusters nodes using a hybrid clustering strategy. By introducing contrastive learning between nodes and their clustering centers, the collaborative semantic information between similar nodes under the target behavior is effectively captured and amplified, alleviating the challenge of sparse target interaction data. Extensive experiments on three real-world datasets show that MBDCC consistently outperforms state-of-the-art baselines, it achieves remarkable improvements of up to 6.84% in Recall@10 and 5.18% in NDCG@50. Moreover, ablation experiments further demonstrate the correctness of our motivation and the necessity of the various modules of the MBDCC model.

People's social relationships are often manifested through their surroundings, with certain objects or interactions acting as symbols for specific relationships, e.g., wedding rings, roses, hugs, or holding hands. This brings unique challenges to recognizing social relationships, requiring understanding and capturing the essence of these contexts from visual appearances. However, current methods of social relationship understanding rely on the basic classification paradigm of detected persons and objects, which fails to understand the comprehensive context and often overlooks decisive social factors, especially subtle visual cues. To highlight the social-aware context and intricate details, we propose a novel approach that recognizes contextual social relationships (ConSoRs) from a social cognitive perspective. Specifically, to incorporate social-aware semantics, we build a lightweight adapter upon the frozen contrastive language-image pretraining (CLIP) to learn social concepts via our novel multimodal side adapter tuning mechanism. Furthermore, we construct social-aware descriptive language prompts (e.g., scene, activity, objects, and emotions) with social relationships for each image, and then compel ConSoR to concentrate more intensively on the decisive visual social factors via visual-linguistic contrasting. Impressively, ConSoR outperforms previous methods with a 7.6% gain on the people-in-social-context (PISC) dataset and a 9.8% increase on the people-in-photo-album (PIPA) benchmark. Furthermore, we observe that ConSoR excels at finding critical visual evidence to reveal social relationships. Code is available at https://github.com/starsholic/ConSoR.

Large Language Models (LLMs) have demonstrated remarkable proficiency in understanding and generating natural language, with an increasing presence in the field of recommendation systems. However, LLMs still encounter a significant issue known as prompt sensitivity, where the model’s output is susceptible to minor changes in the input prompt. This challenge is particularly problematic in recommendation systems, which rely on stable and consistent outputs. Fine-tuning LLMs with diverse prompts can reduce prompt sensitivity but also lead to a decline in recommendation performance. Therefore, choosing an effective fine-tuning method is important to achieve enhanced robustness without sacrificing performance. To address prompt sensitivity while maintaining recommendation performance, we propose Generative Adversarial Network-based prompt enhancement (GANPrompt), a framework for improving LLM-based recommendation systems using adversarial game theory. In this framework, the generator and discriminator compete to produce diverse prompts, which are then used to fine-tune LLM-based recommendation systems, enhancing both robustness and accuracy. Specifically, to generate diverse prompts for fine-tuning and enhance the robustness of LLMs, we develop a GAN-based generator for diverse prompts, with an attribute generation module providing the foundational data support. Further, we introduce a diversity constraint to ensure that the generated prompts maintain high diversity while preserving semantic consistency. To maintain accuracy during the fine-tuning process, we introduce an explicit guidance knowledge token integration method. This method reduces noise and information loss in the face of diverse prompts by enhancing the use of traditional collaborative signals. Through extensive experiments on four publicly available datasets and one real-world industrial dataset, we demonstrate the effectiveness of the proposed framework. Our source code is available at https://github.com/LxytIUON/GANPrompt .

Significant efforts have been dedicated to integrating the powerful Large Language Models (LLMs) with diverse modalities, particularly focusing on the fusion of language, vision and audio data. However, the graph-structured data, which is inherently rich in structural and domain-specific knowledge, has not yet been gracefully adapted to LLMs. Existing methods either describe the graph with raw text, suffering the loss of graph structural information, or feed Graph Neural Network (GNN) embeddings into LLMs at the cost of losing explainable prompt semantics. To bridge this gap, we introduce an end-to-end modality-aligning framework for LLM-graph alignment: Dual-Residual Vector Quantized-Variational AutoEncoder, namely Dr.E. Our approach is purposefully designed to facilitate token-level alignment with LLMs, enabling an effective translation of the intrinsic `language' of graphs into comprehensible natural language. We also manage to enhance LLMs' more robust structural understanding of graphs by incorporating multiple views of the central nodes based on their surrounding nodes at various distances. Our experimental evaluations on standard graph tasks demonstrate competitive performance against other state-of-the-art (SOTA) approaches. Additionally, our framework ensures certain visual interpretability, efficiency, and robustness, marking the promising successful endeavor to achieve token-level alignment between LLMs and GNNs.

Attention mechanisms are critical to the success of large language models (LLMs), driving significant advancements in multiple fields. However, for graph-structured data, which requires emphasis on topological connections, they fall short compared to message-passing mechanisms on fixed links, such as those employed by Graph Neural Networks (GNNs). This raises a question: “Does attention fail for graphs in natural language settings?” Motivated by these observations, we embarked on an empirical study from the perspective of attention mechanisms to explore how LLMs process graph-structured data. The goal is to gain deeper insights into the attention behavior of LLMs over graph structures. Through a series of experiments, we uncovered unique phenomena regarding how LLMs apply attention to graph-structured data and analyzed these findings to improve the modeling of such data by LLMs. The primary findings of our research are: 1) While LLMs can recognize graph data and capture text-node interactions, they struggle to model inter-node relationships within graph structures due to inherent architectural constraints. 2) The attention distribution of LLMs across graph nodes does not align with ideal structural patterns, indicating a failure to adapt to graph topology nuances. 3) Neither fully connected attention (as in LLMs) nor fixed connectivity (as in GNNs) is optimal; each has specific limitations in its application scenarios. Instead, intermediate-state attention windows improve LLM training performance and seamlessly transition to fully connected windows during inference. Source code: https://anonymous.4open.science/r/LLM_exploration-B21F

Although previous studies have applied diffusion models to time series forecasting, these efforts have struggled to preserve the intrinsic temporal correlations within the series, leading to suboptimal predictive outcomes. This failure primarily results from the introduction of independent, identically distributed (i.i.d.) noise. In the forward process, the addition of i.i.d. noise to the time series gradually diminishes these temporal correlations. The reverse process starts with i.i.d. noise and lacks priors related to temporal correlations, which can result in directional biases during sampling. From a frequency-domain perspective, noise disrupts the low-frequency-dominated structure of trend components, making it difficult for the model to learn long-term temporal dependencies. To address these limitations, we introduce a decomposition prediction framework to complement the novel Temporal Correlation-Empowered Diffusion Model. Overall, We decompose the time series into trend and residual components, predict them using a base model and a diffusion model, and then combine the results. Specifically, a frequency-domain MLP model was adopted as the base model due to its not distorting the original sequence, and better the capture of long-range temporal dependencies. The diffusion model incorporates two key modules to capture short- and mid-range temporal correlations: the Maintaining Temporal Correlation Module and the Redesigned Initial Module. Extensive experiments across multiple datasets demonstrate that the proposed method significantly outperforms related strong baselines.

In the realm of recommender systems, accurately predicting Click-Through Rate (CTR) is a critical task that involves learning user-item interaction features. Many researchers propose novel models to mine interaction signals, but they neglect that redundancy itself causes high computational cost and leads to suboptimal performance. Some tried to remove redundancy by dropping useless features, or shrinking the size of embedding table. However, current feature selection methods are vulnerable to training stochasticity and data dynamics, while embedding size assignment techniques neglect the importance relationships between feature fields. The simple combination of the two optimization ways will also yield poor performance due to the inherent gap in their optimization targets. Hence, there is no effective paradigm that can optimize feature fields from the two aspects in a simultaneous and coordinated way. In this paper, we identify the core issue as the lack of a practical score to measure the contribution of feature fields, and propose a distribution-based field optimization framework that adopts importance distribution to provide a comprehensive view for both methods. We innovatively design a learner for each field to acquire the stable and comprehensive importance situation. Then, based on this, we eliminate noise features, and assign adaptive embedding sizes for different feature fields according to the similarity of importance. With this field optimization, our proposed framework has extremely low pre-training overhead, greatly reduces training and inference time, and even achieves more accurate prediction results with fewer feature fields.

Liping Liu, Chunhong Zhang, Likang Wu, Chuang Zhao, Zheng Hu, Ming He, Jianping Fan. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

An intelligent code search engine tries to find and suggest a code piece given a developer's query quickly from a large-scale program database, which can significantly promote software development efficiency. Existing solutions can search the relevant codes to some extent. However, most of them fail to precisely understand the search intent of developers since they only mine their natural language queries, while ignoring the valuable programming context (e.g., the code written by the developer). In this paper, we study the novel problem of context-aware code search. To promote a step forward, we first provide the CodeSearchNet-C dataset with constructing sufficient programming context from the GitHub website for each query-code instance. The dataset is supplemented on the CodeSearchNet benchmark, ensuring both generality and comparability for relevant research. Then, by analyzing the characteristics of programming context, we propose a novel two-stage Context-aware Code Retrieval (ConCR) framework. In the first stage, we propose a Context Walking algorithm, which simulates the programming habits of different developers. The generated programming context could ensure the diversity of search intent among developers. In the second stage, imitating the reading habits of developers, we introduce a novel Context Hierarchical Encoder, to understand the search intent with contextual information from local to global. Our ConCR framework is general, and we give three implementations on the basis of typical code search models as backbones. Extensive experimental results clearly prove that our ConCR significantly enhances the code search performance, effectively fulfilling developers' needs for efficient code resource searching on the web. These results also verify the necessity of introducing programming context to understand developers' intent.

Sequential recommendation, where user preference is dynamically inferred from sequential historical behaviors, is a critical task in recommender systems (RSs). To further optimize long-term user engagement, offline reinforcement-learning-based RSs have become a mainstream technique as they provide an additional advantage in avoiding global explorations that may harm online users’ experiences. However, previous studies mainly focus on discrete action and policy spaces, which might have difficulties in handling dramatically growing items efficiently. To mitigate this issue, in this paper, we aim to design an algorithmic framework applicable to continuous policies. To facilitate the control in the low-dimensional but dense user preference space, we propose an Efficient Continuous Control framework (ECoC). Based on a statistically tested assumption, we first propose the novel unified action representation abstracted from normalized user and item spaces. Then, we develop the corresponding policy evaluation and policy improvement procedures. During this process, strategic exploration and directional control in terms of unified actions are carefully designed and crucial to final recommendation decisions. Moreover, beneficial from unified actions, the conservatism regularization for policies and value functions are combined and perfectly compatible with the continuous framework. The resulting dual regularization ensures the successful offline training of RL-based recommendation policies. Finally, we conduct extensive experiments to validate the effectiveness of our framework. The results show that compared to the discrete baselines, our ECoC is trained far more efficiently. Meanwhile, the final policies outperform baselines in both capturing the offline data and gaining long-term rewards. © 2025 The Authors

Recently, numerous studies have integrated self-supervised contrastive learning with Graph Convolutional Networks (GCNs) to address the data sparsity and popularity bias to enhance recommendation performance. While such studies have made breakthroughs in accuracy metric, they often neglect non-accuracy objectives such as diversity, novelty and percentage of long-tail items, which greatly reduces the user experience in real-world applications. To this end, we propose a novel graph collaborative filtering model named Multi-Objective Graph Contrastive Learning for recommendation (MOGCL), designed to provide more comprehensive recommendations by considering multiple objectives. Specifically, MOGCL comprises three modules: a multi-objective embedding generation module, an embedding fusion module and a transfer learning module. In the multi-objective embedding generation module, we employ two GCN encoders with different goal orientations to generate node embeddings targeting accuracy and non-accuracy objectives, respectively. These embeddings are then effectively fused with complementary weights in the embedding fusion module. In the transfer learning module, we suggest an auxiliary self-supervised task to promote the maximization of the mutual information of the two sets of embeddings, so that the obtained final embeddings are more stable and comprehensive. The experimental results on three real-world datasets show that MOGCL achieves optimal trade-offs between multiple objectives comparing to the state-of-the-arts.

Knowledge discovery and utilization are two essential cognitive processes that enable humans to understand the world and extract new insights from their surroundings. These processes have motivated machine learning studies, particularly zero-shot (ZS) learning, which seeks to identify unseen concepts through the use of side information. Previous ZS studies primarily focused on utilizing existing knowledge to infer unseen events, yet they overlook the crucial process of knowledge discovery and the integrated modeling of these knowledge-aware processes. In this study, we present a comprehensive ZS learning approach that explores and evaluates the machine’s abilities of discovering and utilizing knowledge. More specifically, to emulate human-like knowledge discovery and utilization processes, we propose a novel visual-aware ZS knowledge graph completion task for evaluation, incorporating a traditional ZS image classification task. Technically, we develop a unified ZS learning paradigm named Cognitive Learner (CoLa) to foster the two knowledge-aware abilities. Including a knowledge representation learning (KRL) module and a knowledge adaptation (KA) module, CoLa adapts well to the two specified tasks with the corresponding data. Extensive experiments on large-scale datasets demonstrate CoLa models’ outstanding performance over compared methods in the two ZS tasks, illustrating their superior ability of discovering and utilizing knowledge.

Large Language Models (LLMs) are increasingly prominent in the recommendation systems domain. Existing studies usually utilize in-context learning or supervised finetuning on task-specific data to align LLMs into recommendations. However, the substantial bias in semantic spaces between language processing tasks and recommendation tasks poses a non-negligible challenge. Specifically, without the adequate capturing ability of collaborative information, existing modeling paradigms struggle to capture behavior patterns within community groups, leading to LLMs' ineffectiveness in discerning implicit interaction semantic in recommendation scenarios. To address this, we consider enhancing the learning capability of language modeldriven recommendation models for structured data, specifically by utilizing interaction graphs rich in collaborative semantics. We propose a Graph-Aware Learning for Language ModelDriven Recommendations (GAL-Rec). GAL-Rec enhances the understanding of user-item collaborative semantics by imitating the intent of Graph Neural Networks (GNNs) to aggregate multihop information, thereby fully exploiting the substantial learning capacity of LLMs to independently address the complex graphs in the recommendation system. Sufficient experimental results on three real-world datasets demonstrate that GAL-Rec significantly enhances the comprehension of collaborative semantics, and improves recommendation performance.

Despite advancements using graph neural networks (GNNs) to capture complex user-item interactions, challenges persist due to data sparsity and noise. To address these, self-supervised learning (SSL) methods, particularly recent generative approaches, have gained attention due to their ability to augment graph data without requiring complex view constructions and unstable negative sampling. However, existing generative SSL solutions often focus on structural rather than semantic (refer to collaborative signals in recommendation scenarios) reconstruction, limiting their potential as comprehensive recommender. This paper explores the untapped potential of generative SSL for graph-based recommender systems. We highlight two critical challenges: firstly, designing effective diffusion mechanisms to enhance semantic information and collaborative signals while avoiding optimization biases; and secondly, developing adaptive structural masking mechanisms within graph diffusion to improve overall model performance. Motivated by these challenges, we propose a novel approach: the Guided Diffusion enhanced Mask graph AutoEncoder (GDiffMAE). GDiffMAE integrates an adaptive mask encoder for structural reconstruction and a guided diffusion model for semantic reconstruction, addressing the limitations of current methods. Experimental results on diverse datasets demonstrate that GDiffMAE consistently outperforms powerful baseline models, particularly in handling noisy data scenarios. By enhancing both structural and semantic dimensions through guided diffusion, our model advances the state-of-the-art in graph-based recommender systems.

Recently, the multi-behavior information on a specific domain has been successfully exploited by aggregating diverse user behaviors to solve the problems of cold start and data sparsity in recommendations. However, the user behavior information captured from multiple behaviors in a single domain is insufficient. Our study seeks to enhance user behavior prediction by leveraging both multi-behavior information and cross-domain information in a more effective manner. In order to explore the correlations and differences between different behaviors and different domains, we propose a novel competition framework consists of intra-domain competition and inter-domain competition for knowledge learning. Specifically, for intra-domain, a behavior competition mechanism is designed to enable the model to mine users’ interests and behavior patterns effectively. For inter-domain, a domain competition mechanism is designed to perform knowledge transfer and knowledge fusion for overlapping users in different domains. Through the competition mechanisms, our proposedGraph Competitive Transfer Network (GCTN)achieves knowledge transfer between different domains and captures users’ behavior patterns in different contexts. The effectiveness of the GCTN and its competition mechanisms has been validated through sufficient experimental trials onDoubanandAmazondatasets. Compared to baseline methods, GCTN has demonstrated a marked improvement in both$AUC$and$F1$scores.

The rapidly changing landscape of technology and industries leads to dynamic skill requirements, making it crucial for employees and employers to anticipate such shifts to maintain a competitive edge in the labor market. Existing efforts in this area either relies on domain-expert knowledge or regarding the skill evolution as a simplified time series forecasting problem. However, both approaches overlook the sophisticated relationships among different skills and the inner-connection between skill demand and supply variations. In this paper, we propose a Cross-view Hierarchical Graph learning Hypernetwork (CHGH) framework for joint skill demand-supply prediction. Specifically, CHGH is an encoder-decoder network consisting of i) a cross-view graph encoder to capture the interconnection between skill demand and supply, ii) a hierarchical graph encoder to model the co-evolution of skills from a cluster-wise perspective, and iii) a conditional hyper-decoder to jointly predict demand and supply variations by incorporating historical demand-supply gaps. Extensive experiments on three real-world datasets demonstrate the superiority of the proposed framework compared to seven baselines and the effectiveness of the three modules.

Large Language Models (LLMs) have revolutionized natural language processing tasks, demonstrating their exceptional capabilities in various domains. However, their potential for graph semantic mining in job recommendations remains largely unexplored. This paper focuses on unveiling the capability of large language models in understanding behavior graphs and leveraging this understanding to enhance recommendations in online recruitment, including promoting out-of-distribution (OOD) applications. We present a novel framework that harnesses the rich contextual information and semantic representations provided by large language models to analyze behavior graphs and uncover underlying patterns and relationships. Specifically, we propose a meta-path prompt constructor that aids LLM recommender in grasping the semantics of behavior graphs for the first time and design a corresponding path augmentation module to alleviate the prompt bias introduced by path-based sequence input. By facilitating this capability, our framework enables personalized and accurate job recommendations for individual users. We evaluate the effectiveness of our approach on comprehensive real-world datasets and demonstrate its ability to improve the relevance and quality of recommended results. This research not only sheds light on the untapped potential of large language models but also provides valuable insights for developing advanced recommendation systems in the recruitment market. The findings contribute to the growing field of natural language processing and offer practical implications for enhancing job search experiences.

Zero-Shot Relational Learning (ZSRL), strives to predict relations that have not been observed during training, presenting a considerable challenge in terms of model generalization. Existing ZSRL methods usually utilize the prior knowledge of labels (e.g., text description, ontological schema) to enable knowledge transfer by learned features. Nonetheless, these methods remain limited to calculating the surface features exhibited by relations, failing to fully explore their underlying driving factors. This leads to insufficient discrimination between the shared and distinctive inherent components among relations, which consequently impedes the cognitive understanding required for advanced reasoning. In our study, we aim to identify and utilize shared factors that widely exist in the prior knowledge of classes to learn enhanced semantic representations via shared factors composition, and develop our Factor-based ZSRL framework (FZR) with Generative Adversarial Networks (GANs) to bridge inequality between seen and unseen classes. FZR is designed to restructure the semantic space in such a way that it captures the essence of relation formation, thereby facilitating superior knowledge transfer in zero-shot scenarios. We conduct extensive experiments and evaluate our model on real-world datasets, and the results clearly demonstrate the effectiveness of the proposed model in zero-shot relational learning tasks.

The emergence of personalized generation has made it possible to create texts or images that meet the unique needs of users.Recent advances mainly focus on style or scene transfer based on given keywords.However, in e-commerce and recommender systems, it is almost an untouched area to explore user historical interactions, automatically mine user interests with semantic associations, and create item representations that closely align with user individual interests.In this paper, we propose a brand new framework called Interest Augmented Multimodal Generator (I-AM-G).The framework first extracts tags from the multimodal information of items that the user has interacted with, and the most frequently occurred ones are extracted to rewrite the text description of the item.Then, the framework uses a decoupled text-to-text and image-to-image retriever to search for the top-K similar item text and image embeddings from the item pool.Finally, the Attention module for user interests fuses the retrieved information in a cross-modal manner and further guides the personalized generation process collaborating with the rewritten text.We conducted extensive and comprehensive experiments to demonstrate that our framework can effectively generate results aligned with user preferences, which potentially provides a new paradigm of Rewrite and Retrieve for personalized generation.

Sequential recommendations have gained great attention in recent years. Considering that user interactions naturally follow sequential decision tasks, many researchers have explored the integration of Reinforcement Learning (RL) into sequential recommendations. However, the limited exploration gap between recommender systems and other RL scenarios brings great challenges. Fortunately, Offline RL, which learns from expert experiences in offline datasets, has emerged as a potential solution. Nevertheless, utilizing offline RL to model user preferences is non-trivial due to strict model overestimation and implicit feedback issues. To address these challenges, we propose a novel offline RL framework called Pair-Wise Q-Learning (PQL). Our approach includes a state encoder module to generate a unified representation of user interaction states, a conservative Q-Learning approach with double Q-Networks to reduce overestimation and enhance robustness, a dynamic negative sampling strategy based on Q-values, and a pair-wise learning module to handle implicit feedback. We jointly optimize our model using conservative Q-Learning and pair-wise learning. Experimental results on two real-world e-commerce datasets, considering clicks and purchases, demonstrate the superior performance of our proposed PQL over existing methods.

To alleviate the long-standing data sparsity issue in recommender systems, numerous studies in cross-domain recommendation (CDR) have been conducted to facilitate information transfer processes across domains. In recent years, dual-target CDR has been introduced to gain mutual improvements between two domains through more general bidirectional transfer rather than traditional one-way transit. Existing methods in dual-target CDR focus primarily on designing powerful encoders to learn representative cross-domain information, without tackling the fundamental issue of interaction data shortage. In this paper, we present CrossAug, a novel data augmentation approach to leverage interactions more efficiently in two domains. Specifically, we propose intra-domain and inter-domain augmentations based on cross-reconstructed representations in terms of sampled records. To reduce the harm of domain shift, we project domain-shared representations in two domains into a joint space with Householder transformations and apply center alignments. All these modules boost the utilization of interactions with little influence from negative transfer. Extensive experimental results over public datasets demonstrate the effectiveness of CrossAug and its components in dual-target CDR.

In recommendation systems, users frequently engage in multiple types of behaviors, such as clicking, adding to cart, and purchasing. Multi-behavior sequential recommendation aims to jointly consider multiple behaviors to improve the target behavior's performance. However, with diversified behavior data, user behavior sequences will become very long in the short term, which brings challenges to the efficiency of the sequence recommendation model. Meanwhile, some behavior data will also bring inevitable noise to the modeling of user interests. To address the aforementioned issues, firstly, we develop the Efficient Behavior Sequence Miner (EBM) that efficiently captures intricate patterns in user behavior while maintaining low time complexity and parameter count. Secondly, we design hard and soft denoising modules for different noise types and fully explore the relationship between behaviors and noise. Finally, we introduce a contrastive loss function along with a guided training strategy to contrast the valid information with the noisy signal in the data, and seamlessly integrate the two denoising processes to achieve a high degree of decoupling of the noisy signal. Sufficient experiments on real-world datasets demonstrate the effectiveness and efficiency of our approach in dealing with multi-behavior sequential recommendation.

With the explosive growth of scholarly information, researchers spend much time and effort copiously quoting authoritative works to support their ideas or motivations. We aim to alleviate this situation by proposing a citation recommendation strategy that recalls related papers for a rough idea (a piece of text, i.e., abstract, manuscript). However, the perspective of existing citation recommendations can not be well applied to our task for two defects. First, these methods neglect the reasoning of research topics, which makes the recommendation mechanism not meticulous enough and lacks explainability. For instance, they are not able to mine the hidden citing logic for the candidate paper while recommending. We fill the research gap by constructing structural topics consisting of knowledge concepts from the textual content, where reasoning paths between topics are extracted from an external knowledge graph. Second, the citation network is viewed as a crucial structural context to enhance the recommendation performance, but the new target idea does not have links to the citation network as published papers do. To simulate the prospective topological structure, our model, meanwhile, incorporates a contrastive-learning-based alignment paradigm to encourage the consistency of content embeddings and structure-oriented embeddings. We evaluate our proposed model on three real-world datasets and demonstrate that it significantly improves recommendation accuracy while providing high-quality knowledge-aware reasoning. And an interesting visual example illustrates the reasoning process when our model actually judges samples, which supports the feasibility of our topic-view learning paradigm.

In recent years, multi-behavior information has been utilized to address data sparsity and cold-start issues. The general multi-behavior models capture multiple behaviors of users to make the representation of relevant features more fine-grained and informative. However, most current multi-behavior recommendation methods neglect the exploration of social relations between users. Actually, users’ potential social connections are critical to assist them in filtering multifarious messages, which may be one key for models to tap deeper into users’ interests. Additionally, existing models usually focus on the positive behaviors (e.g., click , follow , and purchase ) of users and tend to ignore the value of negative behaviors (e.g., unfollow and badpost ). In this work, we present a Multi-Behavior Graph (MBG) construction method based on user behaviors and social relationships and then introduce a novel socially enhanced and behavior-aware graph neural network for behavior prediction. Specifically, we propose a Socially Enhanced Heterogeneous Graph Convolutional Network (SHGCN) model, which utilizes behavior heterogeneous graph convolution module and social graph convolution module to effectively incorporate behavior features and social information to achieve precise multi-behavior prediction. In addition, the aggregation pooling mechanism is suggested to integrate the outputs of different graph convolution layers, and a dynamic adaptive loss (DAL) method is presented to explore the weight of each behavior. The experimental results on the datasets of the e-commerce platforms (i.e., Epinions and Ciao) indicate the promising performance of SHGCN. Compared with the most powerful baseline, SHGCN achieves 3.3% and 1.4% uplift in terms of AUC on the Epinions and Ciao datasets. Further experiments, including model efficiency analysis, DAL mechanism, and ablation experiments, confirm the validity of the multi-behavior information and social enhancement.

Federated recommendation (FedRec) can train personalized recommenders without collecting user data, but the decentralized nature makes it susceptible to poisoning attacks. Most previous studies focus on the targeted attack to promote certain items, while the untargeted attack that aims to degrade the overall performance of the FedRec system remains less explored. In fact, untargeted attacks can disrupt the user experience and bring severe financial loss to the service provider. However, existing untargeted attack methods are either inapplicable or ineffective against FedRec systems. In this paper, we delve into the untargeted attack and its defense for FedRec systems. (i) We propose ClusterAttack, a novel untargeted attack method. It uploads poisonous gradients that converge the item embeddings into several dense clusters, which make the recommender generate similar scores for these items in the same cluster and perturb the ranking order. (ii) We propose a uniformity-based defense mechanism (UNION) to protect FedRec systems from such attacks. We design a contrastive learning task that regularizes the item embeddings toward a uniform distribution. Then the server filters out these malicious gradients by estimating the uniformity of updated item embeddings. Experiments on two public datasets show that ClusterAttack can effectively degrade the performance of FedRec systems while circumventing many defense methods, and UNION can improve the resistance of the system against various untargeted attacks, including our ClusterAttack.

Recommendation systems have attracted attention from academia and industry due to their wide range of application scenarios. However, cold start remains a challenging problem limited by sparse user interactions. Some scholars propose to transfer the dense information from the source domain to the target domain through cross-domain recommendation, but most of the work assumes that there is a small amount of historical interaction in the target domain. However, this approach essentially presupposes the existence of at least some historical interaction within the target domain. In this paper, we focus on the domain-level zero-shot recommendation (DZSR) problem. To address the above challenges, we propose a knowledge-aware cross-semantic alignment (K-CSA) framework to learn transferable source domain semantic information. The motivation is to establish stable alignments of interests in different domains through class semantic descriptions (CSDs). Specifically, due to the lack of effective information in the target domain, we learn semantic representations of source and target domain items based on knowledge graphs. Moreover, we conduct multi-view K-means to extract item CSDs from the learned semantic representations. Further, K-CSA learns universal user CSDs through the designed multi-head self-attention. To facilitate the transference of user interest from the source domain to the target domain, we devise a cross-semantic contrastive learning strategy, grounded in the prototype distribution matrix. We conduct extensive experiments on several real-world cross-domain datasets, and the experimental results clearly demonstrate the superiority of our proposed K-CSA compared with other baselines.

The sequential recommendation system has been widely studied for its promising effectiveness in capturing dynamic preferences buried in users' sequential behaviors. Despite the considerable achievements, existing methods usually focus on intra-sequence modeling while overlooking exploiting global collaborative information by inter-sequence modeling, resulting in inferior recommendation performance. Therefore, previous works attempt to tackle this problem with a global collaborative item graph constructed by pre-defined rules. However, these methods neglect two crucial properties when capturing global collaborative information, i.e., adaptiveness and personalization, yielding sub-optimal user representations. To this end, we propose a graph-driven framework, named Adaptive and Personalized Graph Learning for Sequential Recommendation (APGL4SR), that incorporates adaptive and personalized global collaborative information into sequential recommendation systems. Specifically, we first learn an adaptive global graph among all items and capture global collaborative information with it in a self-supervised fashion, whose computational burden can be further alleviated by the proposed SVD-based accelerator. Furthermore, based on the graph, we propose to extract and utilize personalized item correlations in the form of relative positional encoding, which is a highly compatible manner of personalizing the utilization of global collaborative information. Finally, the entire framework is optimized in a multi-task learning paradigm, thus each part of APGL4SR can be mutually reinforced. As a generic framework, APGL4SR can not only outperform other baselines with significant margins, but also exhibit promising versatility, the ability to learn a meaningful global collaborative graph, and the ability to alleviate the dimensional collapse issue of item embeddings.

Recently, Zero-Shot Node Classification (ZNC) has been an emerging and crucial task in graph data analysis. This task aims to predict nodes from unseen classes which are unobserved in the training process. Existing work mainly utilizes Graph Neural Networks (GNNs) to associate features' prototypes and labels' semantics thus enabling knowledge transfer from seen to unseen classes. However, the multi-faceted semantic orientation in the feature-semantic alignment has been neglected by previous work, i.e. the content of a node usually covers diverse topics that are relevant to the semantics of multiple labels. It's necessary to separate and judge the semantic factors that tremendously affect the cognitive ability to improve the generality of models. To this end, we propose a Knowledge-Aware Multi-Faceted framework (KMF) that enhances the richness of label semantics via the extracted KG (Knowledge Graph)-based topics. And then the content of each node is reconstructed to a topic-level representation that offers multi-faceted and fine-grained semantic relevancy to different labels. Due to the particularity of the graph's instance (i.e., node) representation, a novel geometric constraint is developed to alleviate the problem of prototype drift caused by node information aggregation. Finally, we conduct extensive experiments on several public graph datasets and design an application of zero-shot cross-domain recommendation. The quantitative results demonstrate both the effectiveness and generalization of KMF with the comparison of state-of-the-art baselines.

Zero-Shot Learning (ZSL), which aims at automatically recognizing unseen objects, is a promising learning paradigm to understand new real-world knowledge for machines continuously. Recently, the Knowledge Graph (KG) has been proven as an effective scheme for handling the zero-shot task with large-scale and non-attribute data. Prior studies always embed relationships of seen and unseen objects into visual information from existing knowledge graphs to promote the cognitive ability of the unseen data. Actually, real-world knowledge is naturally formed by multimodal facts. Compared with ordinary structural knowledge from a graph perspective, multimodal KG can provide cognitive systems with fine-grained knowledge. For example, the text description and visual content can depict more critical details of a fact than only depending on knowledge triplets. Unfortunately, this multimodal fine-grained knowledge is largely unexploited due to the bottleneck of feature alignment between different modalities. To that end, we propose a multimodal intensive ZSL framework that matches regions of images with corresponding semantic embeddings via a designed dense attention module and self-calibration loss. It makes the semantic transfer process of our ZSL framework learns more differentiated knowledge between entities. Our model also gets rid of the performance limitation of only using rough global features. We conduct extensive experiments and evaluate our model on large-scale real-world data. The experimental results clearly demonstrate the effectiveness of the proposed model in standard zero-shot classification tasks.

Graph Neural Networks (GNNs) such as Graph Convolutional Networks (GCNs) can effectively learn node representations via aggregating neighbors based on the relation graph. However, despite a few exceptions, most of the previous work in this line does not consider the topical semantics underlying the edges, making the node representations less effective and the learned relation between nodes hard to explain. For instance, the current GNNs make us usually don’t know what is the reason for the connection of network nodes, such as the specific research topics cited in this article and the concerns among friends on social platforms. Some methods have begun to explore the extraction of relation semantics in recent related literature, but existing studies generally face two bottlenecks, i.e., either being unable to explain the mined latent relations to ensure their reasonableness and independence, or demanding the textual content of edges which is unavailable in most real-world datasets. Actually, these two issues are both crucial in practical use. In our work, we propose a novel Topic-Disentangled Graph Neural Network (TDG) to address the above two issues at the same time, which explores the relation topics from the perspective of node contents. We design an optimized graph topic module to handle node features to construct independent and explainable semantic subspaces, then the reasonable relation topics that correspond to these subspaces are assigned to each graph relation via a neighborhood routing mechanism. Our proposed model can be easily combined with related graph tasks to form an end-to-end model, to avoid the risk of deviation between node representation space and task space. To evaluate the efficiency of our model, sufficient node-related tasks are conducted on three public datasets in the experimental section. The results show the obvious superiority of TDG compared with the state-of-the-art models.

Network-aware cascade size prediction aims to predict the final reposted number of user-generated information via modeling the propagation process in social networks. Estimating the user's reposting probability by social influence, namely state activation plays an important role in the information diffusion process. Therefore, Graph Neural Networks (GNN), which can simulate the information interaction between nodes, has been proved as an effective scheme to handle this prediction task. However, existing studies including GNN-based models usually neglect a vital factor of user's preference which influences the state activation deeply. To that end, we propose a novel framework to promote cascade size prediction by enhancing the user preference modeling according to three stages, i.e., preference topics generation, preference shift modeling, and social influence activation. Our end-to-end method makes the user activating process of information diffusion more adaptive and accurate. Extensive experiments on two large-scale real-world datasets have clearly demonstrated the effectiveness of our proposed model compared to state-of-the-art baselines.

Representation learning of examination papers is the cornerstone of the Examination Paper Analysis (EPA) in education area including Paper Difficulty Prediction (PDR) and Finding Similar Papers (FSP). Previous works mainly focus on the representation learning of each test item, but few works notice the hierarchical document structure in examination papers. To this end, in this paper, we propose a novel Examination Organization Encoder (EOE) to learn a robust representation of the examination paper with the hierarchical document structure. Specifically, we first propose a syntax parser to recover the hierarchical document structure and convert an examination paper to an Examination Organization Tree (EOT), where the test items are the leaf nodes and the internal nodes are summarization of their child nodes. Then, we applied a two-layer GRU-based module to obtain the representation of each leaf node. After that, we design a subtree encoder module to aggregate the representation of each leaf node, which is used to calculate an embedding for each layer in the EOT. Finally, we feed all the layer embedding into an output module, the process is over and we get the examination paper representation that can be used for downstream tasks. Extensive experiments on real-world data demonstrate the effectiveness and interpretability of our method.

Well begun is half done. In the crowdfunding market, the early fundraising performance of the project is a concerned issue for both creators and platforms. However, estimating the early fundraising performance before the project published is very challenging and still under-explored. To that end, in this paper, we present a focused study on this important problem in a market modeling view. Specifically, we propose a Graph-based Market Environment model (GME) for estimating the early fundraising performance of the target project by exploiting the market environment. In addition, we discriminatively model the market competition and market evolution by designing two graph-based neural network architectures and incorporating them into the joint optimization stage. Finally, we conduct extensive experiments on the real-world crowdfunding data collected from Indiegogo.com. The experimental results clearly demonstrate the effectiveness of our proposed model for modeling and estimating the early fundraising performance of the target project.

Complementary recommendations, which aim at providing users product suggestions that are supplementary and compatible with their obtained items, have become a hot topic in both academia and industry in recent years. Existing work mainly focused on modeling the co-purchased relations between two items, but the compositional associations of item collections are largely unexplored. Actually, when a user chooses the complementary items for the purchased products, it is intuitive that she will consider the visual semantic coherence (such as color collocations, texture compatibilities) in addition to global impressions. Towards this end, in this paper, we propose a novel Content Attentive Neural Network (CANN) to model the comprehensive compositional coherence on both global contents and semantic contents. Specifically, we first propose a Global Coherence Learning (GCL) module based on multi-heads attention to model the global compositional coherence. Then, we generate the semantic-focal representations from different semantic regions and design a Focal Coherence Learning (FCL) module to learn the focal compositional coherence from different semantic-focal representations. Finally, we optimize the CANN in a novel compositional optimization strategy. Extensive experiments on the large-scale real-world data clearly demonstrate the effectiveness of CANN compared with several state-of-the-art methods.