论文检索

Qinxin Zhao, Sheng Zhong

Traditional Intrusion Detection Systems (IDS) are typically trained in specific network environments, and their performance often degrades significantly when deployed in new environments with different attack categories. To address this challenge, we propose and define the task of cross-dataset intrusion detection and design a novel multimodal contrastive learning framework named TriFusion-IDS. This framework represents network traffic from three complementary dimensions: a graph view to capture structural communication patterns, a tabular view to model statistical features, and a textual view to define the semantics of attacks. TriFusion-IDS fuses the graph and tabular representations and aligns them with textual descriptions in a shared embedding space using a CLIP-style contrastive loss function. This semantics-based alignment mechanism enables the model to overcome the effects of zero-shot categories and thus generalize to new network environments. Our extensive experiments on several mainstream datasets demonstrate that this method significantly outperforms existing baselines in cross-dataset intrusion detection scenarios.

Zihan Zhang, Hongzhi Liu, Xiaoshuang Guo, Tianqi Sun, Zhonghai Wu

Medication recommendation systems aim to provide personalized and safe medication options based on individual patient records. However, existing approaches often face challenges related to inadequate modeling of complex relationships within Electronic Health Records (EHRs), data sparsity, and a lack of explainability for recommendations. In this paper, we present a Knowledge-enhanced Explainable HyperGraph Convolution Network (KEHGCN) that constructs a hierarchical hypergraph structure to capture the multi-level relationships within EHR data. By incorporating external knowledge graphs, our approach introduces additional positive relations that help alleviate the impact of data sparsity on model learning. Furthermore, by performing generalized metapath construction and selection on the knowledge graph, our approach achieves effective knowledge filtering and extracts semantically meaningful metapaths, thereby further enhancing the explainability of the recommendation results. We also explicitly introduce negative relations present in the domain knowledge to improve the safety of medication recommendation. Extensive experiments on different hospital departments of MIMIC-III and MIMIC-IV datasets demonstrate that KEHGCN outperforms other state-of-the-art baselines.

Yipeng Zhang, Xin Wang, Hong Chen, Junwei Pan, Qian Li, Jun Zhang, Jie Jiang, Hong Mei, Wenwu Zhu

Sequential recommendation aims to predict the next item based on historical interactions. To further enhance the reasoning capability in sequential recommendation, LLMs are employed to predict the next item or generate semantic IDs for item representation, given LLMs' extensive domain knowledge and reasoning ability. However, existing LLM-based methods suffer from two limitations. (i) The scarcity of recommendation data with reasoning paths makes it challenging to design suitable chain-of-thought prompting templates, and the full potential of LLMs' reasoning abilities remains underutilized. (ii) Upon obtaining semantic IDs, the LLMs and their representations are excluded from the subsequent recommendation model training, preventing downstream models from fully utilizing the rich semantic information encoded within these IDs. To address these issues, we propose a novel CoderRec framework, which is capable of fully exploiting the information encoded in semantic IDs to guide the recommendation process. Specifically, to address the problem of scarcity in reasoning path-augmented data, we introduce latent reasoning into sequential recommendation and treat the representation captured by the downstream model as domain-specific latent thought, enabling implicit logical inference without requiring explicit CoT annotations. To ensure that the downstream recommendation models are able to deeply leverage the semantic information within IDs, we propose a novel cross-scale model collaboration strategy, which employs cross-scale IDs and a two-phase approach to align LLM-derived semantics with recommendation objectives. Extensive experiments have shown the effectiveness of our proposed CoderRec framework.

Yingyi Zhang, Pengyue Jia, Derong Xu, Yi Wen, Xianneng Li, Yichao Wang, Wenlin Zhang, Xiaopeng Li, Weinan Gan, Huifeng Guo 等

Retrieval-Augmented Generation (RAG) critically depends on effective query expansion to retrieve relevant information. However, existing expansion methods adopt uniform strategies that overlook user-specific semantics, ignoring individual expression styles, preferences, and historical context. In practice, identical queries in text can express vastly different intentions across users. This representational rigidity limits the ability of current RAG systems to generalize effectively in personalized settings. Specifically, we identify two core challenges for personalization: 1) user expression styles are inherently diverse, making it difficult for standard expansions to preserve personalized intent. 2) user corpora induce heterogeneous semantic structures—varying in topical focus and lexical organization—which hinders the effective anchoring of expanded queries within the user’s corpora space. To address these challenges, we propose Personalize Before Retrieve (PBR), a framework that incorporates user-specific signals into query expansion prior to retrieval. PBR consists of two components: P-PRF, which generates stylistically aligned pseudo feedback using user history for simulating user expression style, and P-Anchor, which performs graph-based structure alignment over user corpora to capture its structure. Together, they produce personalized query representations tailored for retrieval. Experiments on two personalized benchmarks show that PBR consistently outperforms strong baselines, with up to 10% gains on PersonaBench across retrievers. Our findings demonstrate the value of modeling personalization before retrieval to close the semantic gap in user-adaptive RAG systems.

Yimei Zhang, Guojiang Shen, Kaili Ning, Tongwei Ren, Xuebo Qiu, Mengmeng Wang, Xiangjie Kong

Region representation learning plays a pivotal role in urban computing by extracting meaningful features from unlabeled urban data. Analogous to how perceived facial age reflects an individual's health, the visual appearance of a city serves as its "portrait", encapsulating latent socio-economic and environmental characteristics. Recent studies have explored leveraging Large Language Models (LLMs) to incorporate textual knowledge into imagery-based urban region representation learning. However, two major challenges remain: i) difficulty in aligning fine-grained visual features with long captions, and ii) suboptimal knowledge incorporation due to noise in LLM-generated captions. To address these issues, we propose a novel pre-training framework called UrbanLN that improves Urban region representation learning through Long-text awareness and Noise suppression. Specifically, we introduce an information-preserved stretching interpolation strategy that aligns long captions with fine-grained visual semantics in complex urban scenes. To effectively mine knowledge from LLM-generated captions and filter out noise, we propose a dual-level optimization strategy. At the data level, a multi-model collaboration pipeline automatically generates diverse and reliable captions without human intervention. At the model level, we employ a momentum-based self-distillation mechanism to generate stable pseudo-targets, facilitating robust cross-modal learning under noisy conditions. Extensive experiments across four real-world cities and various downstream tasks demonstrate the superior performance of our UrbanLN.

Ye Zhang, Jinlong He, Dongjie Wang, Yupeng Zhou, Minghao Yin

To cultivate students' aesthetic development, teachers must objectively interpret and evaluate the artistic qualities and emotional resonance within their paintings—a process known as aesthetic perception. This evaluation process is labor-intensive and susceptible to biases due to variations among individual teachers. Advances in artificial intelligence (AI) motivate the use of AI-driven models to automate and enhance this aesthetic perception task. However, building effective AI-driven aesthetic perception models requires extensive datasets, which are typically labor-intensive and costly to gather. To address this, we propose a novel framework that selectively identifies the most challenging dimensions of aesthetic perception for expert annotation, using AI-generated pseudo-annotations to reduce cost and improve model performance. Our framework integrates a multi-agent active learning strategy to systematically annotate scores across multiple dimensions of aesthetic perception. Initially, we train an aesthetic perception model using a small, manually annotated dataset, establishing primary annotation capabilities. Then, this trained model generates pseudo-annotations for unlabeled data across various aesthetic dimensions (e.g., humor, happiness). To ensure annotation quality and relevance, a multi-agent system evaluates these pseudo-annotations, identifying dimensions requiring expert human input based on metrics such as model estimation confidence. Human experts provide targeted annotations selectively, refining the dataset and guiding an iterative improvement cycle. Through repeated refinement, the model progressively enhances both its predictive accuracy and its automated annotation proficiency. Our optimization approach dynamically balances accuracy, annotation relevance, and human effort. Extensive experiments conducted on two real-world datasets demonstrate the effectiveness of our framework.

Xinxun Zhang, Pengfei Jiao, Mengzhou Gao, Tianpeng Li, Xuan Guo

Although dynamic graph neural networks (DyGNNs) have demonstrated promising capabilities, most existing methods ignore out-of-distribution (OOD) shifts that commonly exist in dynamic graphs. Dynamic graph OOD generalization is non-trivial due to the following challenges: 1) Identifying invariant and variant patterns amid complex graph evolution, 2) Capturing the intrinsic evolution rationale from these patterns, and 3) Ensuring model generalization across diverse OOD shifts despite limited data distribution observations. Although several attempts have been made to tackle these challenges, none has successfully addressed all three simultaneously, and they face various limitations in complex OOD scenarios. To solve these issues, we propose a Dynamic graph Causal Invariant Learning (DyCIL) model for OOD generalization via exploiting invariant spatio-temporal patterns from a causal view. Specifically, we first develop a dynamic causal subgraph generator to identify causal dynamic subgraphs explicitly. Next, we design a causal-aware spatio-temporal attention module to extract the intrinsic evolution rationale behind invariant patterns. Finally, we further introduce an adaptive environment generator to capture the underlying dynamics of distributional shifts. Extensive experiments on both real-world and synthetic dynamic graph datasets demonstrate the superiority of our model over state-of-the-art baselines in handling OOD shifts.

Xiao Zhang, Xingyu Zhao, Yuan Cao, Bin Wang, Guiyuan Jiang, Yanwei Yu

With the widespread use of location-tracking technologies, large volumes of trajectory data are continuously generated. Trajectory similarity computation is a core task in trajectory mining with broad applications. However, existing methods still face two key challenges: (1) the difficulty of balancing efficiency and representation quality, and (2) the reliance on a single training paradigm, which limits the ability to capture both pairwise similarity and batch-level coherence. To address these challenges, we propose a trajectory similarity computation framework named TrajAgg. Specifically, our framework incorporates a novel Aggregation Transformer that efficiently aggregates GPS and grid features through two stages of direct interaction and enhances the expressiveness of the resulting trajectory embeddings. In addition, by integrating two distinct training paradigms, our model captures both fine-grained pairwise relationships and global structural consistency. We further analyze its effectiveness from the perspective of mutual information. Extensive experiments on three publicly available datasets show that TrajAgg consistently outperforms state-of-the-art baselines. Our method achieves average improvements of 15.11%, 16.49%, 10.41%, and 40.15% in HR@1 under four distance measures across three datasets, respectively.

Wenkang Zhang, Yan Zhao, Qiang Wang, Zhixin Xu, Li Song, Zhengxue Cheng

Free-Viewpoint Video (FVV) enables immersive 3D experiences, but efficient compression of dynamic 3D representation remains a major challenge. Existing dynamic 3D Gaussian Splatting methods couple reconstruction with optimization-dependent compression and customized motion formats, limiting generalization and standardization. To address this, we propose D-FCGS, a novel Feedforward Compression framework for Dynamic Gaussian Splatting. Key innovations include: (1) a standardized Group-of-Frames (GoF) structure with I-P coding, leveraging sparse control points to extract inter-frame motion tensors; (2) a dual prior-aware entropy model that fuses hyperprior and spatial-temporal priors for accurate rate estimation; (3) a control-point-guided motion compensation mechanism and refinement network to enhance view-consistent fidelity. Trained on Gaussian frames derived from multi-view videos, D-FCGS generalizes across diverse scenes in a zero-shot fashion. Experiments show that it matches the rate-distortion performance of optimization-based methods, achieving over 40 times compression compared to the baseline while preserving visual quality across viewpoints. This work advances feedforward compression of dynamic 3DGS, facilitating scalable FVV transmission and storage for immersive applications.

Wei Zhang, Siyu Yi, Lezhi Chen, Yifan Wang, Ziyue Qiao, Yongdao Zhou, Wei Ju

Spatial transcriptomics (ST) enables joint profiling of gene expression and spatial positions, thereby revealing spatially resolved biological functions. However, many existing ST analysis methods often fail to explicitly quantify the belief and uncertainty in decisions caused by noisy ST data, making it difficult to handle spots of varying quality in a fine-grained manner. In addition, domain identification is a fundamental and critical task in ST, but commonly used models that separate expression learning and clustering often struggle to learn cluster-friendly latent representations effectively. To address these issues, we propose PREST, a prototype-based evidence-aware integration framework for ST data. PREST performs multi-scale representation learning with fine-grained attention fusion and introduces learnable class prototypes to quantify belief and uncertainty in model decisions. We aim to align overall belief scores with latent semantic information to enhance uncertainty quantification and prototype learning, thereby promoting the learning of clustering-friendly representations. PREST further integrates an uncertainty-aware reconstruction module and spatial regularization to reduce overfitting to unreliable spots and promote denoised, discriminative representations. Extensive experiments on several benchmark datasets validate the effectiveness and superiority of our proposed PREST across various downstream tasks.

Shuili Zhang, Hongzhang Mu, Jiawei Sheng, Qianqian Tong, Wenyuan Zhang, Quangang Li, Tingwen Liu

Attribute-specific fashion retrieval aims to enhance fine-grained image retrieval by emphasizing the similarity of specific attributes. Current methods primarily rely on attention mechanisms to extract attribute-related visual features but face two key challenges: the limitations of coarse-grained localization in achieving fine-grained accuracy, and an imbalance between global and local perception, where excessive focus on local features can undermine overall performance. To address these issues, we propose the fashion microscope ProFashion, which achieves pixel-level attribute awareness through optimal transport and neural semantic aggregation. The framework begins by employing optimal transport to align semantic attributes with visual patterns from a global perspective, generating an attribute-visual value map that highlights distinctive regions while reducing interference. This is followed by simulating the human brain's perception of attribute feature patterns through superpixel generation and aggregation, capturing attribute-related features at the pixel semantic level and forming key semantic clusters that preserve microstructures. Building on this, an attribute graph is constructed to facilitate feature clustering, significantly enhancing the framework's capability to handle overlapping features and cross-scale relationships. Comprehensive experiments on the FashionAI, DeepFashion, and DARN datasets demonstrate the framework's effectiveness, achieving overall MAP improvements of 3.11%, 3.70%, and 3.49%, respectively. Additionally, the framework delivers relative average throughput gains of 26.94%, 22.22%, and 24.78% on the FashionAI, DeepFashion, and DARN datasets, respectively.

Jingyuan Zhang, Xin Wang, Lei Yu, Li Yang, Fengjun Zhang

Graph Neural Networks (GNNs) have achieved impressive performance in semi-supervised graph anomaly detection (GAD). While many GNN variants have been developed for this task, they largely focus on advanced message aggregation schemes, leaving the message routing aspect underexplored. We argue that the commonly used broadcast-based routing can also hinder generalization, particularly in the presence of rare and structurally challenging (vertices with a high-degree) anomalies. To address this, we propose Binary Message Passing (BMP), a novel routing paradigm that models the message flow of each vertex as a binary tree (BMP tree), where vanilla graph convolution is decoupled by its left and right subtrees. Each vertex recursively gathers information from neighbors with higher anomaly probabilities within each subtree, thereby amplifying the propagation of anomaly information across the topology. The anomaly probabilities are estimated and updated by the model itself, enabling adaptive, self-supervised routing over iterations. Furthermore, combining multiple BMP trees into a BMP forest provides multi-scale structural context, enhancing the expressiveness of final vertex embeddings. Extensive experiments show that BMP improves detection performance under limited supervision while exhibiting better generalization across structurally diverse anomalies.

Jingmao Zhang, Zhiting Zhao, Yunqi Lin, Jianghong Ma, Tianjun Wei, Haijun Zhang, Xiaofeng Zhang

Beyond user-item modeling, item-to-item relationships are increasingly used to enhance recommendation. However, common methods largely rely on co-occurrence, making them prone to item popularity bias and user attributes, which degrades embedding quality and performance. Meanwhile, although diversity is acknowledged as a key aspect of recommendation quality, existing research offers limited attention to it, with a notable lack of causal perspectives and theoretical grounding. To address these challenges, we propose Cadence: Diversity Recommendation via Causal Deconfounding of Co-purchase Relations and Counterfactual Exposure—a plug-and-play framework built upon LightGCN as the backbone, primarily designed to enhance recommendation diversity while preserving accuracy. First, we compute the Unbiased Asymmetric Co-purchase Relationship (UACR) between items—excluding item popularity and user attributes—to construct a deconfounded directed item graph, with an aggregation mechanism to refine embeddings. Second, we leverage UACR to identify diverse categories of items that exhibit strong causal relevance to a user's interacted items but have not yet been engaged with. We then simulate their behavior under high-exposure scenarios, thereby significantly enhancing recommendation diversity while preserving relevance. Extensive experiments on real-world datasets demonstrate that our method consistently outperforms state-of-the-art diversity models in both diversity and accuracy, and further validates its effectiveness, transferability, and efficiency over baselines.

Jian Zhang, Zhiyuan Wang, Zhangqi Wang, Fangzhi Xu, Qika Lin, Lingling Zhang, Rui Mao, Erik Cambria, Jun Liu

Collaborative reasoning with multiple agents offers the potential for more robust and diverse problem-solving. However, existing approaches often suffer from homogeneous agent behaviors and lack of reflective and rethinking capabilities. We propose Multi-Agent Personality Shaping ((MAPS), a novel framework that enhances reasoning through agent diversity and internal critique. Inspired by the Big Five personality theory, MAPS assigns distinct personality traits to individual agents, shaping their reasoning styles and promoting heterogeneous collaboration. To enable deeper and more adaptive reasoning, MAPS introduces a Critic agent that reflects on intermediate outputs, revisits flawed steps, and guides iterative refinement. This integration of personality-driven agent design and structured collaboration improves both reasoning depth and flexibility. Empirical evaluations across three benchmarks demonstrate the strong performance of MAPS, with further analysis confirming its generalizability across different large language models and validating the benefits of multi-agent collaboration.

Jian Zhang, Zhangqi Wang, Haiping Zhu, Kangda Cheng, Kai He, Bo Li, Qika Lin, Jun Liu, Erik Cambria

Large language models (LLMs) typically operate in a question-answering paradigm, where the quality of the input prompt critically affects the response. Automated Prompt Optimization (APO) aims to overcome the cognitive biases of manually crafted prompts and explore a broader prompt design space. However, existing APO methods often suffer from rigid template structures and inefficient exploration in the prompt space. To this end, we propose a Multi-Agent Adaptive Reasoning with Socratic guidance framework (MARS) for APO. MARS consists of five complementary agents and formulates the optimization process as a Partially Observable Markov Decision Process (POMDP), enabling adaptive prompt refinement through explicit state modeling and interactive feedback. Specifically, a Planner agent generates flexible optimization trajectories, a Teacher-Critic-Student triad engages in Socratic-style dialogue to iteratively optimize the prompt based on pseudo-gradient signals in the text space, and a Target agent executes the prompt in downstream tasks to provide performance feedback. MARS integrates reasoning, feedback, and state transition into a unified hidden-state evolution process, improving both the effectiveness and interpretability of optimization. Extensive experiments on multiple datasets demonstrate that MARS outperforms existing APO methods in terms of optimization performance, search efficiency, and interpretability.

Huizhe Zhang, Jintang Li, Yuchang Zhu, Huazhen Zhong, Liang Chen

Graph Transformers (GTs), which integrate message passing and self-attention mechanisms simultaneously, have achieved promising empirical results in graph prediction tasks. However, the design of scalable and topology-aware node tokenization has lagged behind other modalities. This gap becomes critical as the quadratic complexity of full attention renders them impractical on large-scale graphs. Recently, Spiking Neural Networks (SNNs), as brain-inspired models, provided an energy-saving scheme to convert input intensity into discrete spike-based representations through event-driven spiking neurons. Inspired by these characteristics, we propose a linear-time Graph Transformer with Spiking Node Tokenization (GT-SNT) for node classification. By integrating multi-step feature propagation with SNNs, spiking node tokenization generates compact, locality-aware spike count embeddings as node tokens to avoid predefined codebooks and their utilization issues. The codebook guided self-attention leverages these tokens to perform node-to-token attention for linear-time global context aggregation. In experiments, we compare GT-SNT with other state-of-the-art baselines on node classification datasets ranging from small to large. Experimental results show that GT-SNT achieves comparable performances on most datasets and reaches up to 130× faster inference speed compared to other GTs.

Hongzhi Zhang, Jiameng Chen, Kun Li, Yida Xiong, Xiantao Cai, Wenbin Hu, Jia Wu

The prediction of compound–protein interactions (CPIs) is crucial for drug discovery. Most existing CPI prediction models rely on protein sequence information as input. However, in early-stage drug development, particularly in phenotype-driven studies or compound-response analyses, proteins are often annotated only with functional labels, and their sequences remain undetermined. Consequently, current methods are inapplicable in such scenarios. Furthermore, our experiments find that even when large-scale perturbations were applied to protein sequences, the predictive performance of the existing models did not show a significant decline. It indicates that the high investment in sequencing may not bring corresponding returns. To address the above issues, we propose an inexpensive, protein-sequencing-free framework BioText-CPI, based on the Biomedical Textual description of protein for CPI prediction. Firstly, during the pre-training stage of the model, we use contrastive learning to align protein texts and sequence modalities. Subsequently, we add biological text descriptions of proteins to the existing public CPI dataset to construct a new CPI dataset. Finally, in the CPI prediction stage, the sequence and biomedical text descriptions of proteins can be used as the input for CPI prediction either separately or simultaneously to meet the application requirements of different scenarios. The experiments demonstrate that BioText-CPI achieves comparable effects to the traditional methods when only the biomedical description of protein is input. Moreover, when the two modalities of protein information are input simultaneously, BioText-CPI achieves state-of-the-art performance across multiple scenarios.

Guixian Zhang, Yanmei Zhang, Guan Yuan, Shang Liu, Xiaojing Du, Debo Cheng

Graph Neural Networks (GNNs) have effectively improved the performance of Cognitive Diagnosis Models (CDMs). Existing works have proposed a series of Graph-based Cognitive Diagnosis Frameworks (GCDFs) to enhance robustness to noise. However, these robust designs are often general methods for GNNs and are not designed for cognitive diagnosis, which undermines real cognitive information during the denoising process. Interestingly, a noteworthy phenomenon has been overlooked: even without robustness designs, GCDFs can still learn correct information in noisy environments. In this paper, we conduct a comprehensive empirical analysis of this issue. We found that noise primarily accumulates in lower singular components. Even in noisy environments, the principal subspaces of representations still remain stable. Based on these findings, we propose a Noise-aware Cognitive Diagnostic framework based on Low-rank Alignment, named NCDLA. The framework first performs low-rank reconstruction of the interaction matrix between students and exercises, retaining only larger singular values to achieve noise reduction. Then, the reconstructed interaction matrix and the original interaction matrix are combined with the Q matrix to form a noise-reduced heterogeneous graph and an original heterogeneous graph. In order to distinguish between the interaction patterns of correct and incorrect responses, we decompose the heterogeneous graph according to the type of response. NCDLA achieves denoising of student representations and exercises representations through a self-supervised strategy based on low-rank reconstruction and a spectral anchor regularisation method. Extensive experiments on three datasets demonstrate that NCDLA achieves optimal prediction performance and robustness.

Fuwei Zhang, Xiaoyu Liu, Dongbo Xi, Jishen Yin, Huan Chen, Peng Yan, Fuzhen Zhuang, Zhao Zhang

Generative Recommendation (GR) has emerged as a new paradigm in recommender systems. This approach relies on quantized representations to discretize item features, modeling users’ historical interactions as sequences of discrete tokens. Based on these tokenized sequences, GR predicts the next item by employing next-token prediction methods. The challenges of GR lie in constructing high-quality semantic identifiers (IDs) that are hierarchically organized, minimally conflicting, and conducive to effective generative model training. However, current approaches remain limited in their ability to harness multimodal information and to capture the deep and intricate interactions among diverse modalities, both of which are essential for learning high-quality semantic IDs and for effectively training GR models. To address this, we propose Multi-Aspect Cross-modal quantization for generative Recommendation (MACRec), which introduces multimodal information and incorporates it into both semantic ID learning and generative model training from different aspects. Specifically, we first introduce cross-modal quantization during the ID learning process, which effectively reduces conflict rates and thus improves codebook usability through the complementary integration of multimodal information. In addition, to further enhance the generative ability of our GR model, we incorporate multi-aspect cross-modal alignments, including the implicit and explicit alignments. Finally, we conduct extensive experiments on three well-known recommendation datasets to demonstrate the effectiveness of our proposed method.

Fan Zhang, Jinpeng Chen, Tao Wang, Huan Li, Senzhang Wang, Feifei Kou, Ye Ji, Kaimin Wei, Zhenye Yang

Out-of-Town (OOT) recommendation aims to provide personalized suggestions for users in unfamiliar cities. However, OOT recommendation faces two fundamental challenges: the difficulty of reasoning across modalities, as preference signals in disparate formats such as images and text are hard to compare; and the preference deviation problem, since a user's resident and tourist preferences often diverge, rendering simple preference transfer ineffective. To address these challenges, we propose Distinguishing Resident and Tourist Preferences via Multi-Modal LLM Alignment for Out-of-Town Cross-Domain Recommendation (DiMA), a framework for re-ranking Points of Interest (POIs). To tackle the multimodal challenge, DiMA first leverages Multimodal Large Language Models and Large Language Models (LLMs) to transform heterogeneous POI data into unified semantic tags, enabling both cross-modal reasoning and efficient downstream processing. To address preference deviation, a ``teacher'' LLM executes a custom Chain-of-Thought (CoT) process to disentangle resident and tourist preferences from multi-city histories for re-ranking. Finally, a lightweight student model learns this CoT reasoning via Supervised Fine-Tuning and is then refined with Direct Preference Optimization to align with true user choices, with the potential to surpass the teacher. Extensive experiments on a real-world dataset demonstrate that DiMA significantly enhances the performance of baseline models in the OOT recommendation re-ranking task.