Traditional recommendation algorithms primarily focus on accuracy, often overlooking diversity and long-tail item recommendations, which are inherently conflicting objectives. Balancing these objectives remains a significant challenge. To address this issue, we propose refined objective functions for diversity and long-tail item recommendation, aiming to more accurately evaluate the diversity of recommendation lists and improve the exposure of long-tail items. To solve the resulting multi-objective optimization problem, we introduce a preference-guided multi-objective recommendation algorithm incorporating an adaptive Tchebycheff environmental selection strategy. This algorithm combines a preference-guided local search to enhance solution quality with an adaptive Tchebycheff selection mechanism that dynamically adjusts objective weights to improve both convergence and stability. Experimental results demonstrate that the proposed algorithm effectively balances accuracy, diversity, and long-tail item recommendation.

This work introduces a Structurally-Aware Contextual Distillation framework for entity completion in Knowledge Graphs to tackle the neglect of graph structure and directional bias. The approach integrates a Multi-Hop Neighborhood Generation module using Graph Neural Networks and a Reverse Triple Description module. Notably, using a compact Vicuna7B model for distillation, it achieves competitive performance against the original work that relied on a 540 B parameter teacher. Experiments on WN18RR and FB15k-237N demonstrate improvements in Mean Reciprocal Rank and Hits@k, confirming enhanced structural awareness and a more balanced performance in head versus tail entity prediction.

Multimodal sentiment analysis (MSA) integrates and processes data from multiple sources, like audio and text, to better understand human emotions through cross-modal interactions. Previous research generally obtain a feature embedding from different modalities and fuse at the utterance level directly, making it difficult to capture fine-grained multimodal representations and susceptible to irrelevant information from heterogeneous modalities. Moreover, these methods often overlook explicit word-level alignment between modalities, which limits effective representation learning and multimodal fusion. Therefore, we propose a novel method, Fine-grained Multimodal Fusion Network(MMTA) for MSA. Specifically, a fine-grained representation alignment module is introduced to extract representation at the word level by montreal forced aligner (MFA). A local-global multi-scale fusion method is then designed to further capture more fine-grained interaction from local features, i.e., word-level features. Finally, we also investigate the effectiveness of incorporating word-level contextual information into the fine-grained fusion process. Extensive experiments on the public MSA datasets, i.e., MOSI and MOSEI, show that our approach surpass previous baselines.

Recommender systems based on collaborative filtering (CF) effectively model user-item interactions but struggle with data sparsity and cold-start issues. On the other hand, large language models (LLMs) offer strong semantic understanding but fail to capture structured user-item relationships. To bridge this gap, we propose AlignGenRec, a framework that integrates collaborative and textual knowledge for generative recommendation. Specifically, we introduce an embedding alignment mechanism that aligns item embeddings from a pre-trained CF model with text embeddings from item descriptions. These aligned embeddings are then transferred to the LLM without requiring fine-tuning, enabling structured knowledge integration while preserving generative capabilities. Additionally, constrained sequence decoding ensures that generated recommendations correspond to valid items, improving recommendation accuracy. Experimental results demonstrate that AlignGenRec outperforms both CF-based and LLM-based baselines, particularly in cold-start scenarios, with performance improvements of 7 - 11% across three benchmark datasets. Beyond recommendation tasks, AlignGenRec also supports preference prediction and user profiling, highlighting its versatility in real-world applications.