Multimodal recommendation improves user modeling by integrating collaborative signals with heterogeneous item content. In real applications, user interests evolve over time and exhibit nonstationary dynamics, where different preference factors change at different rates. This challenge is amplified in multimodal settings because visual and textual cues can dominate decisions under different temporal regimes. Despite strong progress, most multimodal recommenders still rely on static interaction graphs or coarse temporal heuristics, which limits their ability to model continuous preference evolution with fine-grained temporal adaptation. To address these limitations, we propose TimeMM, a time-conditioned spectral filtering framework for dynamic multimodal recommendation. TimeMM instantiates Time-as-Operator by mapping interaction recency to a family of parametric temporal kernels that reweight edges on the user--item graph, producing component-specific representations without explicit eigendecomposition. To capture non-stationary interests, we introduce Adaptive Spectral Filtering that mixes the operator bank according to temporal context, yielding prediction-specific effective spectral responses. To account for modality-specific temporal sensitivity, we further propose Spectral-Aware Modality Routing that calibrates visual and textual contributions conditioned on the same temporal context. Finally, a ranking-space Spectral Diversity Regularization encourages complementary expert behaviors and prevents filter-bank collapse. Extensive experiments on real-world benchmarks demonstrate that TimeMM consistently outperforms state-of-the-art multimodal recommenders while maintaining linear-time scalability.

Multimodal recommendation aims to integrate collaborative signals with heterogeneous content such as visual and textual information, but remains challenged by modality-specific noise, semantic inconsistency, and unstable propagation over user–item graphs. These issues are often exacerbated by naive fusion or shallow modeling strategies, leading to degraded generalization and poor robustness. While recent work has explored the frequency domain as a lens to separate stable from noisy signals, most methods rely on static filtering or reweighting, lacking the ability to reason over spectral structure or adapt to modality-specific reliability. To address these challenges, we propose a Structured Spectral Reasoning (SSR) framework for frequency-aware multimodal recommendation. Our method follows a four-stage pipeline: (i) Decompose graph-based multimodal signals into spectral bands via graph-guided transformations to isolate semantic granularity; (ii) Modulate band-level reliability with spectral band masking, a training-time masking with representation-consistency objective that suppresses brittle frequency components; (iii) Fuse complementary frequency cues using hyperspectral reasoning with low-rank cross-band interaction; and (iv) Align modality-specific spectral features via contrastive regularization to promote semantic and structural consistency. Experiments on three real-world benchmarks show consistent gains over strong baselines, particularly under sparse and cold-start settings. Additional analyses indicate that structured spectral modeling improves robustness and provides clearer diagnostics of how different bands contribute to performance.

Large Language Models (LLMs) have garnered significant attention in Recommendation Systems (RS) due to their extensive world knowledge and robust reasoning capabilities. However, a critical challenge lies in enabling LLMs to effectively comprehend and extract insights from massive user behaviors. Current approaches that directly leverage LLMs for user interest learning face limitations in handling long sequential behaviors, effectively extracting interest, and applying interest in practical scenarios. To address these issues, we propose a Hierarchical Tree Search-based User Lifelong Behavior Modeling framework (HiT-LBM). HiT-LBM integrates Chunked User Behavior Extraction (CUBE) and Hierarchical Tree Search for Interest (HTS) to capture diverse interests and interest evolution of user. CUBE divides user lifelong behaviors into multiple chunks and learns the interest and interest evolution within each chunk in a cascading manner. HTS generates candidate interests through hierarchical expansion and searches for the optimal interest with process rating model to ensure information gain for each behavior chunk. Additionally, we design Temporal-Ware Interest Fusion (TIF) to integrate interests from multiple behavior chunks, constructing a comprehensive representation of user lifelong interests. The representation can be embedded into any recommendation model to enhance performance. Extensive experiments demonstrate the effectiveness of our approach, showing that it surpasses state-of-the-art methods. We also deploy HiT-LBM on Kuaishou's online advertising platform, showing 3.5% increase of revenue. We release the implementation code https://github.com/xiayu-cell/HiT-LBM.

Large Language Models (LLMs) are increasingly employed in zero-shot documents ranking, yielding commendable results. However, several significant challenges still persist in LLMs for ranking: (1) LLMs are constrained by limited input length, precluding them from processing a large number of documents simultaneously; (2) The output document sequence is influenced by the input order of documents, resulting in inconsistent ranking outcomes; (3) Achieving a balance between cost and ranking performance is challenging. To tackle these issues, we introduce a novel documents ranking method called TourRank1. which is inspired by the sport tournaments, such as FIFA World Cup. Specifically, we 1) overcome the limitation in input length and reduce the ranking latency by incorporating a multi-stage grouping strategy similar to the parallel group stage of sport tournaments; 2) improve the ranking performance and robustness to input orders by using a points system to ensemble multiple ranking results. We test TourRank with different LLMs on the TREC DL datasets and the BEIR benchmark. The experimental results demonstrate that TourRank delivers state-of-the-art performance at a modest cost.

The task of text semantic matching focuses on measuring the semantic similarity between two texts and is widely applied in search and ranking scenarios. In recent years, pre-trained foundation models based on the Transformer architecture have demonstrated powerful semantic representation capabilities. The pipeline of fine-tuning pre-trained foundation models on downstream semantic matching tasks has achieved promising results and widespread adoption. However, practical downstream scenarios often face severe challenges in terms of data quality and quantity. Ensuring high-quality and large quantities of samples is often difficult. Current research on enhancing pre-trained models for few-shot text semantic matching tasks is still not advanced enough. Therefore, this paper focuses on providing a general enhancement scheme for few-shot text semantic matching tasks. Specifically, we propose an Enhancing Transformer-based Semantic Matching method for few-shot learning through weakly contrastive pre-training, which is named as SEMFormer. First, starting from the token-level and structural-level perspectives, we design a simple and low-cost data augmentation method to construct weakly supervised samples. Then, we use the global semantic representations to construct a contrastive objective from the relation-aspect perspective. Next, we design a contrastive objective based on the alignment-aspect, aiming to achieve effective semantic matching by optimizing the bidirectional semantic awareness between texts. We conducted comprehensive experiments based on five Chinese and English datasets. The experimental results validated that our proposed weakly contrastive pre-training augmentation method significantly improves model performance. Further experiments confirmed the effectiveness of our design. The source code is available at: https://github.com/llm-ml/SEMFormer.

Whether it is an e-commerce platform or a short video platform, the effective use of multi-modal data plays an important role in the recommendation system. More and more researchers are exploring how to effectively use multimodal signals to entice more users to buy goods or watch short videos. Some studies have added multimodal features as side information to the model and achieved certain results. In practice, the purchase behavior of users mainly depends on some personalized intentions. However, it is difficult for neural networks to process noise information and extract high-level intention information effectively. To investigate the benefits of latent intentions and leverage them effectively for recommendation, we propose a Multimodal-aware Multi-intention Learning method for recommendation (MMIL). Specifically, we first construct a multi-intention recommendation framework based on the probability distribution relationship among predicted objectives and intentions, while avoiding intention overfitting. We then design an intention representation learning module to learn accurate multiple intention representations based on intention prototypes. Further, we propose a multi-modal intention perceptron module to learn multi-modal intention representations. In addition, we design self-supervised intention-based and representation-based contrastive objectives to achieve cross-modal representation alignment. On three real-world data sets, the proposed MMIL method outperforms other advanced techniques. The effectiveness of intention modeling and intention alignment is verified by comprehensive experiments. The source code is available at: https://github.com/ml-mindset/MMIL.

Recommendation systems have been widely used in e-commerce, news media, and short video platforms. With the abundance of images, text, and audio information, users often engage in personalized interactions based on their multimodal preferences. With the continuous expansion of application scenarios, cross domain recommendation issues have become important, such as recommendations in both the public and private domains of e-commerce. The current cross domain recommendation methods have achieved certain results through methods such as shared encoders and contrastive learning. However, few studies have focused on the effective extraction and utilization of multimodal information in cross domain recommendations. Furthermore, due to the existence of distribution drift issues, directly constructing feature alignment between source domain and target domain representations is not an effective way. Therefore, we propose a Multimodal Optimal Transport Knowledge Distillation (MOTKD) method for cross domain recommendation. Specifically, we propose a multimodal graph attention network to model the multimodal preference representation of users. Then, we introduce a proxy distribution space as a bridge between the source and target domains. Based on the common proxy distribution, we utilize the optimal transport method to achieve cross domain knowledge transfer. Further, in order to improve the auxiliary training effect of source domain supervised signals on target domain, we design a multi-level cross domain knowledge distillation module. We conducted extensive experiments on two pairs of cross domain datasets composed of four datasets. The experimental results indicate that our proposed MOTKD method outperforms other state-of-the-art models.

Multimodal recommendation system has been widely used in short video platform, e-commerce platform and news media. Multimodal data contains information such as product image and product text, which is often used as auxiliary signal to improve the effect of recommendation system significantly. In order to alleviate the problems of data sparsity and noise, some researchers construct data augmentation to use self-supervised learning to help model training. These methods have achieved certain results. However, most of the work is based on data augmentation in random ways, such as random masking and random perturbation. This random method is likely to lose important information and introduce new noise, resulting in biased augmentation data. Therefore, we propose a Modal-aware Bias Constrained Contrastive Learning method (BCCL) to solve the above problems. Specifically, BCCL introduces a bias-constrained data augmentation method to ensure the quality of augmentation samples. Then the multi-modal semantic information is modeled by the designed modal awareness module. Furthermore, we propose a information alignment module to improve the sparse modal feature learning of the model. We conducted a comprehensive experiment on three real-world data sets, and the experimental results showed that the proposed BCCL outperformed all the state-of-art methods. In-depth experiments have verified the effectiveness of our proposed modules.

Real recommendation systems contain various features, which are often high-dimensional, sparse, and difficult to learn effectively. In addition to numerical features, user reviews contain rich semantic information including user preferences, which are used as auxiliary features by researchers. The methods of supplementing data features based on reviews have certain effects. However, most of them simply concatenate review representations and other features together, without considering that the text representation contains a lot of noise information. In addition, the important intentions contained in user reviews are not modeled effectively. In order to solve the above problems, we propose a novel Review-based Multi-intention Contrastive Learning (RMCL) method. In detail, RMCL proposes an intention representation method based on mixed Gaussian distribution hypothesis. Further, RMCL adopts a multi-intention contrastive strategy, which establishes a fine-grained connection between user reviews and item reviews. Extensive experiments on five real-world datasets demonstrate significant improvements of our proposed RMCL model over the state-of-the-art methods.

In ball sports, such as basketball, the coach can guide players to better offend and defend from a holistic perspective to win the game. Inspired by such scenarios, we introduce a coach-like concept into the decision-making process of cooperative multi-agent systems. We propose a new framework Commander-Soldiers Reinforcement Learning (CSRL), for Multi-Agent systems. Specifically, we introduce a virtual role, Commander, which can obtain and encode global information every T steps and send the encoded global guidance to Soldiers (real agents). Furthermore, we propose Policy Guidance Network (PGN), which can customize the encoded global guidance from Commander based on observations for each Soldier, providing each Soldier with specified guidance to the decision-making process. The Soldier takes into account not only the local action-observation histories but also the specified guidance from PGN when making decisions. We validate CSRL on the challenging StarCraft II micromanagement benchmark, proving that our approach can take advantage of intermittent global information to improve collaborative performance.