Graph Convolutional Networks (GCNs) are predominantly tailored for graphs displaying homophily, where similar nodes connect, but often fail on heterophilic graphs. The strategy of adopting distinct approaches to learn from homophilic and heterophilic components in node-level tasks has been widely discussed and proven effective both theoretically and experimentally. However, in graph-level tasks, research on this topic remains notably scarce. Addressing this gap, our research conducts an analysis on graphs with nodes' category ID available, distinguishing intra-category and inter-category components as embodiment of homophily and heterophily, respectively. We find while GCNs excel at extracting information within categories, they frequently capture noise from inter-category components. Consequently, it is crucial to employ distinct learning strategies for intra- and inter-category elements. To alleviate this problem, we separately learn the intra- and inter-category parts by a combination of an intra-category convolution (IntraNet) and an inter-category high-pass graph convolution (InterNet). Our IntraNet is supported by sophisticated graph preprocessing steps and a novel category-based graph readout function. For the InterNet, we utilize a high-pass filter to amplify the node disparities, enhancing the recognition of details in the high-frequency components. The proposed approach, DivGNN, combines the IntraNet and InterNet with a gated mechanism and substantially improves classification performance on graph-level tasks, surpassing traditional GNN baselines in effectiveness.
论文检索
输入标题、作者或关键词，从 844 篇学术成果中精准定位
AI-based automatic visual inspection systems have been extensively researched to streamline various industrial products' labor-intensive anomaly detection processes. Despite significant advancements, detecting logical anomalies remains challenging due to the multitude of rules governing the assembly of multiple components to create a normal product. Existing methods have relied solely on image information for anomaly detection, resulting in limited accuracy as they fail to account for these diverse complex rules. Instead, humans detect anomalies by comparing the image with pre-defined logic which can be clearly expressed with natural language. Inspired by the human decision process, we propose a logical anomaly detection model that leverages text-based logic like human reasoning. With user-defined rules (i.e., positive rules) and logically distinct negative rules, we train the model using component-aware contrastive learning that increases the similarity between images and positive rules while decreasing the similarity with negative rules. However, accurately comparing textual and visual features is challenging due to multiple components, each governed by different rules, within a single image. To address this, we developed a zero-shot related region detection technique, which guides the model's focus on components relevant to each rule. We evaluated the proposed model on three public datasets and achieved state-of-the-art results in a few-shot logical anomaly detection task. Our findings highlight the potential of integrating vision-language models to enhance logical anomaly detection and utilizing text-based logic in complex industrial settings.
Few-shot learning is crucial for tabular data, where the high cost of annotation often limits the availability of labeled samples. Despite its importance in real-world applications such as healthcare and finance, few-shot learning in tabular domains has received limited attention. To address this, we introduce range-limited augmentation, a novel augmentation strategy for contrastive learning that perturbs numerical features within predefined feature-specific ranges. Unlike conventional augmentations, which may result in false positive pairs during contrastive learning, our approach ensures semantic consistency by restricting augmentations to ranges. A quantitative analysis confirms that range-limited augmentation better preserves task-relevant information compared to existing augmentation techniques. Additionally, we propose FeSTa (Few-Shot Tabular classification benchmark), the first large-scale benchmark designed to systematically evaluate few-shot learning methods in tabular data. FeSTa includes 50 datasets and 32 algorithms spanning supervised, unsupervised, self-supervised, semi-supervised, and foundation models. Experiments on FeSTa show that range-limited augmentation consistently ranks among the top methods, achieving an average rank of 2.6 out of 32 in 1-shot classification, despite not relying on large-scale pretraining or complex architectures. The benchmark code is available in https://github.com/kyungeun-lee/festa.git.
Message Passing Neural Networks (MPNNs) hold a key position in machine learning on graphs, but they struggle with unintended behaviors, such as over-smoothing and over-squashing, due to irregular data structures. The observation and formulation of these limitations have become foundational in constructing more informative graph representations. In this paper, we delve into the limitations of MPNNs, focusing on aspects that have previously been overlooked. Our observations reveal that even within a single layer, the information specific to an individual node can become significantly diluted. To delve into this phenomenon in depth, we present the concept of Over-dilution and formulate it with two dilution factors: intra-node dilution for attribute-level and inter-node dilution for node-level representations. We also introduce a transformer-based solution that alleviates over-dilution and complements existing node embedding methods like MPNNs. Our findings provide new insights and contribute to the development of informative representations. The implementation and supplementary materials are publicly available at https://github.com/LeeJunHyun/NATR.
Traditional multimodal learning approaches often assume that all modalities are available during both the training and inference phases. However, this assumption is often impractical in real-world scenarios due to challenges such as sensor failures, data corruption, or privacy concerns. While recent efforts focus on enhancing the robustness of pre-trained Multimodal Transformers (MTs) under missing modality conditions, mainstream work in this field often overlook reconstructing the missing modalities and rely on static, sample-agnostic prompt-tuning techniques, undermining their efficacy in severe modality missing scenarios. To address these limitations, we propose REDEEM, a novel REtrieval-guiDEd conditional gEnerative fraMework that largely alleviates the modality missing problems on pre-trained MTs. REDEEM consists of a new adaptive retrieval mechanism to identify relevant instances for both modality-complete and -incomplete samples. It then conditions on the remaining modalities and utilizes the retrieved data as experts to effectively recover the missing ones in modality-incomplete instances through a within-modal reconstruction manner. Finally, REDEEM generates sample-aware inter-modal prompts from the retrieved instances to guide MTs in tackling severe modality missing challenges. Comprehensive experiments on three diverse multimodal classification benchmarks demonstrate that REDEEM significantly outperforms competitive baselines.
Despite plentiful successes achieved by graph representation learning in various domains, the training of graph neural networks (GNNs) still remains tenaciously challenging due to the tremendous computational overhead needed for sizable graphs in practice. Recently, graph data distillation (GDD), which seeks to distill large graphs into compact and informative ones, has emerged as a promising technique to enable efficient GNN training. However, most existing GDD works rely on heuristics that align model gradients or representation distributions on condensed and original graphs, leading to compromised result quality, expensive training for distilling large graphs, or both. Motivated by this, this paper presents an efficient and effective GDD approach, ClustGDD. Under the hood, ClustGDD resorts to synthesizing the condensed graph and node attributes through fast and theoretically-grounded clustering that minimizes the within-cluster sum of squares and maximizes the homophily on the original graph. The fundamental idea is inspired by our empirical and theoretical findings unveiling the connection between clustering and empirical condensation quality using Fréchet Inception Distance, a well-known quality metric for synthetic images. Furthermore, to mitigate the adverse effects caused by the homophily-based clustering, ClustGDD refines the nodal attributes of the condensed graph with a small augmentation learned via class-aware graph sampling and consistency loss. Our extensive experiments exhibit that GNNs trained over condensed graphs output by ClustGDD consistently achieve superior or comparable performance to state-of-the-art GDD methods in terms of node classification on five benchmark datasets, while being orders of magnitude faster.
Cross-validation (CV) approaches are widely used in machine learning model development. They select hyperparameters associated with models with high average performance across multiple folds sampled from training data. Popular CV approaches do this sampling either randomly (random CV) or temporally (block CV). Averaging performance across multiple folds is intended to result in robust models; however, it can mask a model's reliance on unstable correlations - transient relationships between inputs and outputs that do not hold over time - which can make a model fail over time. In this paper, we show, both theoretically and empirically, that random and block CV often select hyperparameters corresponding to models that rely on unstable correlations over stable correlations. In light of this shortcoming, we propose a novel CV approach that is more likely to result in models that rely on stable correlations over unstable correlations, compared to random and block CV. Our proposed approach leverages the fact that when a model relies on features whose relationship with the outcome is stable, random and block CV result in the same estimated model performance; it does not matter how you split the training data into CV folds. However, when a model relies on features whose relationship with the outcome is unstable, random and block CV result in different estimates of model performance. We use this difference as a signal that a model relies on an unstable correlation and propose a model selection technique that selects hyperparameters that minimize this difference. When combined with standard model performance criteria, our approach results in models with higher average and more stable performance over time compared to random and block CV. Applied to the task of predicting 5-year survival among individuals with lung cancer, our approach leads to models that perform better in the long term compared to the next best approach (0.850, IQR: [0.824, 0.853] vs. 0.804, IQR: [0.753, 0.850]). In high-stakes dynamic domains like healthcare, where good performance over time is crucial, our approach can help result in models that perform consistently well despite unstable correlations in the training set. Code to implement our approach and reproduce all experiments in the paper is available at https://github.com/mld3/cv_unstable_correlations.
In supervised learning, decision trees are valued for their interpretability and performance. While greedy decision tree algorithms like CART remain widely used due to their computational efficiency, they often produce sub-optimal solutions with respect to a regularized training loss. Conversely, optimal decision tree methods can find better solutions but are computationally intensive and typically limited to shallow trees or binary features. We present Dynamic Programming Decision Trees (DPDT), a framework that bridges the gap between greedy and optimal approaches. DPDT relies on a Markov Decision Process formulation combined with heuristic split generation to construct near-optimal decision trees with significantly reduced computational complexity. Our approach dynamically limits the set of admissible splits at each node while directly optimizing the tree regularized training loss. Theoretical analysis demonstrates that DPDT can minimize regularized training losses at least as well as CART. Our empirical study shows on multiple datasets that DPDT achieves near-optimal loss with orders of magnitude fewer operations than existing optimal solvers. More importantly, extensive benchmarking suggests statistically significant improvements of DPDT over both CART and optimal decision trees in terms of generalization to unseen data. We demonstrate DPDT practicality through applications to boosting, where it consistently outperforms baselines. Our framework provides a promising direction for developing efficient, near-optimal decision tree algorithms that scale to practical applications.
Large language models (LLMs) provide powerful means to leverage prior knowledge for predictive modeling when data is limited. In this work, we demonstrate how LLMs can use their compressed world knowledge to generate intrinsically interpretable machine learning models, i.e., decision trees, without any training data. We find that these zero-shot decision trees can even surpass data-driven trees on some small-sized tabular datasets and that embeddings derived from these trees perform better than data-driven tree-based embeddings on average. Our decision tree induction and embedding approaches can therefore serve as new knowledge-driven baselines for data-driven machine learning methods in the low-data regime. Furthermore, they offer ways to harness the rich world knowledge within LLMs for tabular machine learning tasks. Our code and results are available at https://github.com/ml-lab-htw/llm-trees.
Beyond using multimodal features of items in addition to user-item interactions, researchers have additionally utilized Contrastive Learning (CL) in recent multimedia recommender systems to highly alleviate the data sparsity problem. CL-based methods generate at least two embeddings (i.e., views) for each instance and enrich the information of each instance from various perspectives via the views, thereby alleviating the data sparsity problem. Therefore, CL-based methods have focused on generating views that effectively represent the characteristics of each instance for their downstream tasks. Similarly, CL-based multimedia recommender systems have made efforts to effectively generate their user/item views by leveraging items' multimodal features. However, we point out the following two limitations that they have overlooked in generating their views: (1) they either have not attempted to identify the influence of each modality feature of an item on user-item interactions, or have identified it by randomly masking or dropping user-item interactions, and (2) they have not attempted to identify non-interactions likely to result in interactions in the future. To overcome these limitations, we propose a novel multimedia recommendation framework, named STARLINE, utilizing contraSTive leARning with modaLIty-aware graph refiNEment. Extensive experiments on five real-world datasets validate the effectiveness and validity of STARLINE, especially showing consistently higher accuracy by up to 13.24% compared to the best competitor.
Understanding misalignments in human task-solving trajectories is crucial for enhancing AI models trained to closely mimic human reasoning. This study categorizes such misalignments into three types: (1) Lack of functions to express intent(2) inefficient action sequences, and (3) incorrect intentions that cannot solve the task. To address these issues, we first formalize and define these three misalignment types in a unified framework. We then propose a heuristic algorithm to detect misalignments in ARCTraj trajectories and analyze their impact hierarchically and quantitatively. We also present an intention estimation method based on our formalism that infers missing alignment between user actions and intentions. Through trajectory alignment, we experimentally demonstrate that AI models trained on human task-solving trajectories improve performance in mimicking human reasoning. Based on hierarchical analysis and experiments, we highlight the importance of trajectory-intention alignment and demonstrate the effectiveness of intention-aligned training.
Large Language Models (LLMs) have recently emerged as promising tools for recommendation thanks to their advanced textual understanding ability and context-awareness. Despite the current practice of training and evaluating LLM-based recommendation (LLM4Rec) models under a sequential recommendation scenario, we found that whether these models understand the sequential information inherent in users' item interaction sequences has been largely overlooked. In this paper, we first demonstrate through a series of experiments that existing LLM4Rec models do not fully capture sequential information both during training and inference. Then, we propose a simple yet effective LLM-based sequential recommender, called LLM-SRec, a method that enhances the integration of sequential information into LLMs by distilling the user representations extracted from a pre-trained CF-SRec model into LLMs. Our extensive experiments show that LLM-SRec enhances LLMs' ability to understand users' item interaction sequences, ultimately leading to improved recommendation performance. Furthermore, unlike existing LLM4Rec models that require fine-tuning of LLMs, LLM-SRec achieves state-of-the-art performance by training only a few lightweight MLPs, highlighting its practicality in real-world applications. Our code is available at https://github.com/Sein-Kim/LLM-SRec.
How can we classify graph-structured data with labels of only positive classes? Graph-based MPU learning is to train a classifier where only a few nodes of positive classes are labeled and the others remain unlabeled; i.e., negative labels are completely absent. This scenario is deeply connected to real-world problems such as classifying multiple types of cyberattackers (e.g., DDoS, phishing) in network graphs where explicit labels for normal users are unavailable since the undetected cyberattackers disguise themselves as normal users. The main challenge lies in the connections between nodes of different classes, which cause the learned features of positive and unlabeled nodes to become indistinguishable. This issue is particularly severe for negative nodes as there are no observed labels to guide them in MPU settings. In this paper, we propose D-MVP (Disentangled Multi-View Propagation), an accurate method for graph-based MPU learning. D-MVP disentangles feature propagation into multiple views by assigning distinct weights to each view and aggregating information differently based on these weights. This makes the node classifier learn information that distinguishes between multiple positive classes while simultaneously capturing shared features among them, thereby effectively identifying negative classes that lack these shared characteristics. Extensive experiments on real-world datasets show that D-MVP achieves the best performance.
Diffusion bridges have shown potential in paired image-to-image (I2I) translation tasks. However, existing methods are limited by their unidirectional nature, requiring separate models for forward and reverse translations. This not only doubles the computational cost but also restricts their practicality. In this work, we introduce the Bidirectional Diffusion Bridge Model (BDBM), a scalable approach that facilitates bidirectional translation between two coupled distributions using a single network. BDBM leverages the Chapman-Kolmogorov Equation for bridges, enabling it to model data distribution shifts across timesteps in both forward and backward directions by exploiting the interchangeability of the initial and target timesteps within this framework. Notably, when the marginal distribution given endpoints is Gaussian, BDBM's transition kernels in both directions possess analytical forms, allowing for efficient learning with a single network. We demonstrate the connection between BDBM and existing bridge methods, such as Doob's h-transform and variational approaches, and highlight its advantages. Extensive experiments on high-resolution I2I translation tasks demonstrate that BDBM not only enables bidirectional translation with minimal additional cost but also outperforms state-of-the-art bridge models. Our source code is available at https://github.com/kvmduc/BDBM.
This paper introduces the Lottery Hedge Fund Hypothesis (LHFH) and proposes Knowledge-Aligned Domain Adaptation (KADA), a novel framework for addressing domain discrepancies between source and target domains. While LLMs demonstrate strong generalization across tasks, their performance often degrades on domain-specific tasks requiring specialized knowledge due to a mismatch in domain-specific knowledge distributions. KADA is motivated by LHFH, which posits that models can be viewed as a mixture of knowledge-specialized subnetworks, where domain differences correspond to the weights of these subnetworks. In contrast to conventional domain adaptation methods such as adapters or PEFT, which statically adjust model parameters, this perspective leads to a dynamic adaptation approach: selectively reweighting and activating relevant subnetworks aligned with the target domain. To realize this, KADA introduces two key components: the Knowledge Steering Layer (KSL) - a lightweight, dynamic module that routes input representations through expert subnetworks - and Knowledge Distribution Modeling (KDM) - a training objective that aligns source and target knowledge distributions by regulating routing entropy and balance. These components enable KADA to fine-tune only the subnetworks and routing parameters, avoiding full model retraining and significantly reducing computational overhead. Empirical results demonstrate that KADA effectively bridges domain gaps, achieving competitive or superior performance compared to baseline PEFT methods, even in low-resource settings. KADA offers a scalable and theoretically grounded solution for efficient domain adaptation in LLMs.
We propose One-Shot Multi-view Visual Conversational Recommendation (OSMVR), a novel framework to tackle key challenges in personalized recommendation systems from minimal user input. OSMVR addresses three critical issues: the Cold Start problem, the lack of transparent justifications, and the difficulty of accurately estimating nuanced user preferences. Unlike multi-turn systems, OSMVR reduces user burden by structuring the recommendation process into two phases-Preference Estimation and Recommendation Generation-within a single conversational round. It leverages a multi-view approach that enables one-shot preference inference and multi-perspective recommendation generation. To support this, we introduce View Effective Fine-Tuning (VEFT), a novel method that enhances the reasoning capabilities of Large Language Models (LLMs) and Vision-Language Models (VLMs) for one-shot, multi-view preference inference and recommendation. VEFT dynamically reconfigures subnetworks and soft prompts to simulate expert roles and enable perspective-aware reasoning, capturing diverse aspects of user intent. This lightweight, one shot process can be executed on client devices, ensuring privacy preserving personalization. In the Preference Estimation phase, users provide minimal input, such as text or images of preferred items. OSMVR, using VEFT, infers user preferences without historical data and presents these preferences from multiple viewpoints, inviting user feedback. In the subsequent Recommendation Generation phase, OSMVR generates multi-view recommendations with expert-informed, transparent justifications derived from the inferred preferences. Experimental results show that VEFT significantly improves one shot preference estimation and justification quality, even with limited data. OSMVR proves to be a robust and promising framework for personalized, multimodal recommendation systems, offering a solution that reduces user effort while enhancing satisfaction.
Explaining graphs and their target Graph Neural Networks (GNNs) has gained attention with the growing use of GNNs. Most existing explainable AI (XAI) methods for GNNs focus on extracting an explanation subgraph and assume the target GNN is supervised with accessible class probabilities. However, the growing prevalence of GNN models in unsupervised settings underscores the necessity for task-irrelevant explanations. Moreover, most existing studies scarcely explore whether identifying edges absent from the original graph can improve explanation quality. To this end, we propose HINT-G (Harnessing INfluence function for Task-irrelevant explanation on Graph neural networks), a method that uses influence functions to explain models across diverse learning paradigms and considers edges beyond the given graph. The influence of an edge can be determined directly or by aggregating the influence scores of its constituent nodes, while the influence of a non-existent edge can also be determined. Furthermore, this method is task-irrelevant, since the influence score can be obtained whenever the loss function of the target model is differentiable. Experimental results on several datasets consistently demonstrate that HINT-G effectively explains graphs through the influence function framework. Our implementation code is available at https://github.com/cycy-kim/HINT-G.
Sequential recommendation is a popular paradigm in modern recommender systems. In particular, one challenging problem in this space is cross-domain sequential recommendation (CDSR), which aims to predict future behaviors given user interactions across multiple domains. Existing CDSR frameworks are mostly built on the self-attention transformer and seek to improve by explicitly injecting additional domain-specific components (e.g. domain-aware module blocks). While these additional components help, we argue they overlook the core self-attention module already present in the transformer, a naturally powerful tool to learn correlations among behaviors. In this work, we aim to improve the CDSR performance for simple models from a novel perspective of enhancing the self-attention. Specifically, we introduce a Pareto-optimal self-attention and formulate the cross-domain learning as a multi-objective problem, where we optimize the recommendation task while dynamically minimizing the cross-domain attention scores. Our approach automates knowledge transfer in CDSR (dubbed as AutoCDSR) - it not only mitigates negative transfer but also encourages complementary knowledge exchange among auxiliary domains. Based on the idea, we further introduce AutoCDSR+, a more performant variant with slight additional cost. Our proposal is easy to implement and works as a plug-and-play module that can be incorporated into existing transformer-based recommenders. Besides flexibility, it is practical to deploy because it brings little extra computational overheads without heavy hyper-parameter tuning. We conduct experiments over both large-scale production recommender data as well as academic benchmarks, where AutoCDSR consistently enhances the performance of base transformers, enabling simple models to perform on par with state-of-the-art with less overhead (e.g., 4x faster than state-of-the-art CDSR models). AutoCDSR on average improves Recall@10 for SASRec and Bert4Rec by 9.8% and 16.0% and NDCG@10 by 12.0% and 16.7%, respectively. Code is available at https://github.com/snap-research/AutoCDSR.
Black-box hard-label adversarial attack on text can be used to evaluate the robustness of text classification models (TCMs), allowing for a more thorough examination of potential security flaws before deployment. Research on this problem is still in the embryonic stage and only a few methods are available. Existing adversarial attack methods often treat each adversarial text generation as an independent task, failing to leverage past adversarial experience to facilitate subsequent generation. Consequently, these methods suffer from high query costs (i.e., the number of times TCMs need to be accessed to generate adversarial text) and pose challenges in generating adversarial texts with high semantic similarity and low perturbation rates under a query-limited setup. To achieve query-efficient adversarial text generation, we propose EGAttack, an experience-guided black-box hard-label attack method based on reinforcement learning. Specifically, we models adversarial text generation as a sequential decision process and propose a new adversarial environment to capture the intricate relationships among word perturbations, text semantic, and the target model's outputs for agent learning. Meanwhile, we extend the classical reinforcement framework DDQN into adversarial DDQN (ADDQN) through designing a new adversarial agent and a new dual-objective constraint, driving EGAttack to learn from past adversarial experience while identifying the vulnerabilities of TCMs, thus generating high-quality adversarial texts. Experimental results across various target models and datasets demonstrate that EGAttack generates adversarial texts with higher semantic similarity, lower perturbation rates, and competitive attack success rate, making it an efficient and effective tool for evaluating the robustness of TCMs.
Urban region representation is crucial for various urban downstream tasks. However, despite the proliferation of methods and their success, acquiring general urban region knowledge and adapting to different tasks remains challenging. Existing work pays limited attention to the fine-grained functional layout semantics in urban regions, limiting their ability to capture transferable knowledge across regions. Further, inadequate handling of the unique features and relationships required for different downstream tasks may also hinder effective task adaptation. In this paper, we propose a Graph-based Urban Region Pre-training and Prompting framework (GURPP) for region representation learning. Specifically, we first construct an urban region graph and develop a subgraph-centric urban region pre-training model to capture the heterogeneous and transferable patterns of entity interactions. This model pre-trains knowledge-rich region embeddings using contrastive learning and multi-view learning methods. To further refine these representations, we design two graph-based prompting methods: a manually-defined prompt to incorporate explicit task knowledge and a task-learnable prompt to discover hidden knowledge, which enhances the adaptability of these embeddings to different tasks. Extensive experiments on various urban region prediction tasks and different cities demonstrate the superior performance of our framework.