Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Huaping LiuSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works162 from public data
- Graph Representation Learning via Graphical Mutual Information Maximization706
An unsupervised learning model trained by maximizing GMI between the input and output of a graph neural encoder is developed, which outperforms state-of-the-art unsuper supervised counterparts, and even sometimes exceeds the performance of supervised ones.
- 332
- ANOMALOUS: A Joint Modeling Approach for Anomaly Detection on Attributed Networks264
This paper proposes a novel joint framework to conduct attribute selection and anomaly detection as a whole based on CUR decomposition and residual analysis and results on both synthetic and real-world datasets corroborate the effectiveness of the proposed framework.
- TwiBot-20: A Comprehensive Twitter Bot Detection Benchmark249
TwiBot-20 is a massive Twitter bot detection benchmark, which contains 229,573 users, 33,488,192 tweets, 8,723,736 user property items and 455,958 follow relationships, and is considered to be the largest Twitterbot detection benchmark to date.
- 223
- TwiBot-22: Towards Graph-Based Twitter Bot Detection198
TwiBot-22 is proposed, a comprehensive graph-based Twitter bot detection benchmark that presents the largest dataset to date, provides diversified entities and relations on the Twitter network, and has considerably better annotation quality than existing datasets.
- BotRGCN196
The proposed BotRGCN, which is short for Bot detection with Relational Graph Convolutional Networks, addresses the challenge of community by constructing a heterogeneous graph from follow relationships and applies relational graph convolutional networks.
- Heterogeneity-Aware Twitter Bot Detection with Relational Graph Transformers167
A novel bot detection framework is proposed, which leverages the topological structure of user-formed heterogeneous graphs and models varying influence intensity between users, and proposes relational graph transformers to model heterogeneous influence between users and learn node representations.
- Self-weighted Robust LDA for Multiclass Classification with Edge Classes123
A novel self-weighted robust LDA with ℓ2,1-norm based pairwise between-class distance criterion, called SWRLDA, is proposed, which can automatically avoid the optimal mean calculation and simultaneously learn adaptive weights for each class pair without setting any additional parameter.
- Semantics-Guided Contrastive Network for Zero-Shot Object Detection115
A novel Semantics-Guided Contrastive Network for ZSD is developed, named ContrastZSD, a detection framework that first brings contrastive learning mechanism into the realm of zero-shot detection and outperforms the previous state-of-the-art on both ZSD and generalized ZSD tasks.
- SATAR109
SATAR is proposed, a self-supervised representation learning framework of Twitter users, and applied to the task of bot detection, which demonstrates that SATAR outperforms competitive baselines on different bot detection datasets of varying information completeness and collection time.
- 105
- Scalable attack on graph data by injecting vicious nodes103
This paper develops a more scalable framework named Approximate Fast Gradient Sign Method which considers a more practical attack scenario where adversaries can only inject new vicious nodes to the graph while having no control over the original graph.
- 91
- Graph Few-shot Learning with Attribute Matching84
The proposed AMM-GNN leverages an attribute-level attention mechanism to capture the distinct information of each task and thus learns more effective transferable knowledge for meta-learning.
- 80
- BotMoE: Twitter Bot Detection with Community-Aware Mixtures of Modal-Specific Experts79
BotMoE is proposed, a Twitter bot detection framework that jointly utilizes multiple user information modalities (metadata, textual content, network structure) to improve the detection of deceptive bots and incorporates a community-aware Mixture-of-Experts (MoE) layer to improve domain generalization and adapt to different Twitter communities.
- Simple to complex cross-modal learning to rank79
This paper involves the self-paced learning theory with diversity into the cross-modal learning to rank and learns an optimal multi- modal embedding space based on non-linear mapping functions that enhances the model’s robustness to outliers and achieves better generalization via training the model gradually from easy rankings by diverse queries to more complex ones.
- A new self-supervised task on graphs: Geodesic distance prediction70
This paper introduces a novel self-supervised strategy for graph representation learning by exploiting natural supervision provided by the data itself by randomly selecting pairs of nodes in a graph and training a well-designed neural net to predict the contextual position of one node relative to the other.
- Noisy Correspondence Learning with Meta Similarity Correction69
A Meta Similarity Correction Network (MSCN) is proposed to provide reliable similarity scores and an effective data purification strategy using meta-data as prior knowledge to remove the noisy samples.
- Event-Radar: Event-driven Multi-View Learning for Multimodal Fake News Detection68
An Event-Radar based on multi-view learning based on multi-view learning is proposed based on multi-view learning, which integrates visual manipulation, textual emotion and mul-014 timodal inconsistency at event-level for fake news detection.
- Semantics-Preserving Graph Propagation for Zero-Shot Object Detection68
This paper proposes a novel Semantics-Preserving Graph Propagation model for ZSD based on Graph Convolutional Networks (GCN), and develops a graph construction module to flexibly build category graphs by leveraging diverse correlations between category nodes, followed by two semantics-preserving graph propagation modules that enhance both category and region representations.
- Mitigating World Biases: A Multimodal Multi-View Debiasing Framework for Fake News Video Detection65
A Multimodal Multi-View Debiasing (MMVD) framework, which makes the first attempt to mitigate various multimodal biases for fake news video detection and put forward a multi-view causal reasoning strategy to learn unbiased dependencies within the cognitive biases, thus enhancing the unbiased prediction of multimodal videos.
- 63
- KRACL: Contrastive Learning with Graph Context Modeling for Sparse Knowledge Graph Completion59
This work proposes a novel framework KRACL1 to alleviate the widespread sparsity in KGs with graph context and contrastive learning, and proposes the Knowledge Relational Attention Network (KRAT) to leverage the graph context by simultaneously projecting neighboring triples to different latent spaces and jointly aggregating messages with the attention mechanism.
- 52
- Few-shot activity recognition with cross-modal memory network52
This paper proposes a novel end-to-end cross-modal memory network for few-shot activity recognition that stores the dynamic visual and textual semantics for some high-level attributes related to human activities and could achieve significant improvements over other previous methods.
- 42
- Learning to Rematch Mismatched Pairs for Robust Cross-Modal Retrieval40
The potential semantic similarity among unpaired samples makes it possible to excavate useful knowledge from mismatched pairs and the proposed L2RM, a general framework based on Optimal Transport that learns to rematch mismatched pairs is proposed.
- LMBot: Distilling Graph Knowledge into Language Model for Graph-less Deployment in Twitter Bot Detection39
A novel bot detection framework LMBot is proposed that distills the graph knowledge into language models (LMs) for graph-less deployment in Twitter bot detection to combat data dependency challenge and is compatible with graph-based and graph-less datasets.
- Diverse fuzzy c-means for image clustering39
This work proposes a diverse fuzzy c-means for image clustering via introducing a novel diversity regularization into the traditional fuzzy c -means objective, which guarantees the learned image cluster centers to be different from each other and to fill the image data space as much as possible.
- Probabilistic Non-Negative Matrix Factorization and Its Robust Extensions for Topic Modeling38
This paper follows the generative procedure of topic model and learns the topic-word distribution and topics distribution via directly approximating the word-document co-occurrence matrix with matrix decomposition technique and adopts the explicit probabilistic meaning of factors in topic models.
- 37
- Scheduled sampling for one-shot learning via matching network37
A novel metric is proposed to evaluate the learning difficulty of support set, where this metric jointly considers the semantic diversity and similarity of visual labels and introduces a scheduled sampling strategy to train the matching network from easy to difficult.
- 36
- 35
- 35
- ZSTAD: Zero-Shot Temporal Activity Detection34
This work designs an end-to-end deep network based on R-C3D that is optimized with an innovative loss function that considers the embeddings of activity labels and their super-classes while learning the common semantics of seen and unseen activities.
- Distributed extreme learning machine with alternating direction method of multiplier34
A novel distributed extreme learning machine is exploited to implement theextreme learning machine algorithm in parallel for large-scale data set and a corresponding distributed algorithm is developed on the basis of alternating direction method of multipliers which shows effectiveness in distributed convex optimization.
- 33
- 31
- 30
- Towards Entity Alignment in the Open World: An Unsupervised Approach29
This work develops an unsupervised framework that performs entity alignment in the open world that does not require labeled data and can effectively filter out unmatchable entities.
- 28
- 28
- Noise-Tolerant Learning for Audio-Visual Action Recognition27
A noise-tolerant learning framework that significantly improves the robustness of the action recognition model and surpasses the baselines by a clear margin and establishes a benchmark of real-world noisy correspondence in audio-visual data by relabeling the Kinetics dataset.
- Top-k multi-class SVM using multiple features27
To fuse multiple features effectively, an adaptive weight for each view is introduced and an efficient alternating optimization algorithm is exploited to learn the optimal classifiers and their corresponding weights jointly.
- From Predictions to Analyses: Rationale-Augmented Fake News Detection with Large Vision-Language Models25
This work proposes the Explainable Adaptive Rationale-Augmented Multimodal (EARAM) framework, which adaptively uses MSLMs to extract useful rationales from the multi-perspective analyses of LVLMs, which assists LVLMs in generating more reliable explanations.
- Dual-stream generative adversarial networks for distributionally robust zero-shot learning24
A dual-stream generator to synthesize visual samples and reconstruct semantic embedding simultaneously, thereby ensuring semantic consistency and outperforms the state-of-the-arts by 4.7% and 3.0% on average in two metrics over four real-world datasets.
- An anomaly detection framework for time-evolving attributed networks23
A novel dynamic anomaly detection framework on time-evolving attributed networks on the basis of residual analysis, namely AMAD, which leverages the small smooth disturbance between two consecutive time stamps to characterize the evolution of networks for incrementally update.
- 21
- BotPercent: Estimating Bot Populations in Twitter Communities21
The results reveal that the presence of Twitter bots is not homogeneous, but exhibiting a spatial-temporal distribution with considerable heterogeneity that should be taken into account for content moderation and social media policy making.
- HOFA: Twitter Bot Detection with Homophily-Oriented Augmentation and Frequency Adaptive Attention21
HBOFA is proposed, a novel graph-based Twitter bot detection framework that combats the heterophilous disguise challenge with a homophily-oriented graph augmentation module (Homo-Aug) and a frequency adaptive attention module (FaAt).
- 21
- Adaptive Semi-Supervised Learning with Discriminative Least Squares Regression21
This paper develops a new discriminative LSR by equipping each label with an adjustment vector and distinguishes the label fitting of labeled and unlabeled training data through a probabilistic vector with an adaptive parameter, which always ensures the significant importance of labeled data and characterizes the contribution of unlabeling instance according to its uncertainty.
- Avoiding Optimal Mean ℓ2,1-Norm Maximization-Based Robust PCA for Reconstruction21
This letter equivalently reformulate the objective of conventional PCA and learn the optimal projection directions by maximizing the sum of projected difference between each pair of instances based on -norm by exploiting an efficient optimization algorithm to soften the contributions from outliers by reweighting each data point iteratively.
- 20
- Understand, Refine and Summarize: Multi-View Knowledge Progressive Enhancement Learning for Fake News Video Detection19
This work introduces EvidSV, the first comprehensive benchmark supporting evidence- and attribution-aware fake news video detection, and proposes MUKE, a multi-view knowledge progressive enhancement learning framework that outperforms existing methods in both fake news detection and attribution, and generalizes effectively to previously unseen domains.
- Generating Action-conditioned Prompts for Open-vocabulary Video Action Recognition19
It is argued that augmenting text embeddings with human prior knowledge is pivotal for open-vocabulary video action recognition, and innovatively blend video models with Large Language Models (LLMs) to devise Action-conditioned Prompts.
- 18
- 18
- Robust dictionary learning with graph regularization for unsupervised person re-identification18
A novel unsupervised method, termed as Robust Dictionary Learning with Graph Regularization (RDLGR), which can guarantee view-invariance through learning a dictionary shared by all the camera views, and introduces a cross-view graph Laplacian regularization term into the framework of dictionary learning.
- 17
- Deep Semisupervised Zero-Shot Learning with Maximum Mean Discrepancy17
This work proposes a novel deep semisupervised method that replaces the original labels with the corresponding textual descriptions to better capture the category semantics and overcomes the problem of distribution difference by minimizing the maximum mean discrepancy between seen and unseen instance distributions.
- A novel T–S fuzzy systems identification with block structured sparse representation17
In this paper, accurate description of T–S fuzzy model is established with far fewer numbers of fuzzy rules by selecting the important fuzzy rules and eliminating the redundant ones in the process of block structured sparse regression.
- A new algorithm for testing diagnosability of fuzzy discrete event systems17
This paper investigates the diagnosability of fuzzy discrete event systems by proposing a new algorithm based on the concept of undistinguishable strings, which is obtained in terms of certain properties of the diagnoser which is constructed with respect to the minimal observable event.
- 15
- Disentangled Representation Learning With Transmitted Information Bottleneck15
This work proposes DisTIB (Transmitted Information Bottleneck for Disentangled representation learning), a novel objective that navigates the balance between information compression and preservation and theoretically proves that DisTIB can achieve optimal disentanglement.
- Toward Enhanced Robustness in Unsupervised Graph Representation Learning: A Graph Information Bottleneck Perspective14
This work proposes a novel unbiased robust UGRL method called Robust Graph Information Bottleneck (RGIB), which attempts to learn robust node representations against adversarial perturbations by preserving the original information in the benign graph while eliminating the adversarial information in the adversarial graph.
- Toward Entity Alignment in the Open World: An Unsupervised Approach with Confidence Modeling14
An unsupervised framework UEA is offered that performs entity alignment in the open world and the concept of confidence is introduced to measure the probability of an entity pair of being true and developed a confidence-based un supervised EA framework CUEA.
- Sparse Multi-Modal Topical Coding for Image Annotation14
This paper proposes a non-probabilistic formulation of PTM for automatic image annotation, namely Sparse Multi-Modal Topical Coding, and introduces sparsity to PTM to improve the interpretability of the inferred latent representations.
- ChatGen: Automatic Text-to-Image Generation From FreeStyle Chatting12
This paper proposes Automatic T2I generation, which aims to automate these tedious steps, allowing users to simply describe their needs in a freestyle chatting way, and proposes ChatGen-Evo, a multi-stage evolution strategy that progressively equips models with essential automation skills.
- AutoGPS: Automated Geometry Problem Solving via Multimodal Formalization and Deductive Reasoning12
This work proposes AutoGPS, a neuro-symbolic collaborative framework that solves geometry problems with concise, reliable, and human-interpretable reasoning processes, and achieves state-of-the-art performance on benchmark datasets.
- 12
- 12
- 12
- From Manipulation to Mistrust: Explaining Diverse Micro-Video Misinformation for Robust Debunking in the Wild11
WildFakeBench, a large-scale benchmark of over 10,000 real-world micro-videos covering diverse misinformation types and sources, is introduced, and FakeAgent, a Delphi-inspired multi-agent reasoning framework that integrates multimodal understanding with external evidence for attribution-grounded analysis is developed.
- 11
- 10
- 10
- Empower Post-hoc Graph Explanations with Information Bottleneck: A Pre-training and Fine-tuning Perspective10
A two-stage explanation strategy is considered that endows post-hoc graph explanations with the applicability to pretrained GNNs where downstream tasks are inaccessible and the capacity to explain the transferability of task-agnostic explanations.
- Dynamic T‐S Fuzzy Systems Identification Based on Sparse Regularization10
A dynamic algorithm is proposed to address T‐S fuzzy system identification with both sparsity and dynamic clustering, named as dynamic sparse fuzzy inference systems (D‐sparseFIS) due to two different estimation approaches of fuzzy rule consequence parameter, i.e., global estimation and local estimation.
- PaCo-RL: Advancing Reinforcement Learning for Consistent Image Generation with Pairwise Reward Modeling9
Extensive experiments show that PaCo-Reward significantly improves alignment with human perceptions of visual consistency, and PaCo-GRPO achieves state-of-the-art consistency performance with improved training efficiency and stability, highlighting the promise of PaCo-RL as a practical and scalable solution for consistent image generation.
- GADY: Unsupervised Anomaly Detection on Dynamic Graphs9
This work proposes a continuous dynamic graph model to capture the fine-grained information, which breaks the limit of existing discrete methods, and pioneer the use of Generative Adversarial Networks to generate negative interactions.
- Adversarial Attacks on Fairness of Graph Neural Networks9
This paper proposes G-FairAttack, a general framework for attacking various types of fairness-aware GNNs in terms of fairness with an unnoticeable effect on prediction utility and proposes a fast computation technique to reduce the time complexity of G-FairAttack.
- CGUA: Context-Guided and Unpaired-Assisted Weakly Supervised Person Search9
A novel C ontext-G uided and CGC algorithm to leverage context information in the clustering process and an U npaired- A ssisted M emory ( UAM) unit to distinguish unpaired and paired persons by pushing them away are proposed.
- Reliable shot identification for complex event detection via visual-semantic embedding9
This paper decomposes the video into several segments and intuitively model the task of complex event detection as a multiple instance learning problem by representing each video as a "bag" of segments in which each segment is referred to as an instance.
- 9
- Progressive generative adversarial networks with reliable sample identification9
It is shown by inspection that the reliable samples with smaller errors are beneficial to achieve a better generator, while the unreliable one might disturb the training procedure, and an indicator for each sample to indicate its reliability is introduced in this paper.
- PSDiff: Diffusion Model for Person Search With Iterative and Collaborative Refinement8
A novel Person Search framework based on the Diffusion model, PSDiff, which formulates the person search as a dual denoising process from noisy boxes and ReID embeddings to ground truths and discards prior pedestrian candidates generated by detectors, thereby avoiding the local optimum problem of the ReID task.
- 8
- Nonlinear feature selection on attributed networks8
This paper proposes a novel semi-supervised nonlinear framework FS-GCN based on graph convolutional networks (GCNs) to select high-quality features, which can elaborately catch the nonlinear dependency between nodal attributes and class labels and explores the possibility of totally removing the label information.
- Bridging Interests and Truth: Towards Mitigating Fake News with Personalized and Truthful Recommendations7
Comprehensive evaluations from multiple dimensions demonstrate the superiority of the PRISM model, which incorporates pre-trained authenticity representations as constraints during content generation, ensuring the credibility of the recommended news and effectively curbing the spread of fake news.
- Disentangled Noisy Correspondence Learning7
Inspired by the robustness of information bottlenecks against noise, DisNCL is introduced, a novel information-theoretic framework for feature Disentanglement in Noisy Correspondence Learning to adaptively balance the extraction of modality-invariant information (MII) and MEI with certifiable optimal cross-modal disentanglement efficacy.
- 7
- 6
- 6
- 6
- GraTO6
GraTO is proposed, a framework based on neural architecture search to automatically search for GNNs architecture that demonstrates that GraTo outperforms baselines in the over-smoothing metrics and achieves competitive performance in accuracy.
- 6
- Anomaly Detection in Time-Evolving Attributed Networks6
A novel dynamic anomaly detection framework based on residual analysis, namely AMAD, that leverages the small smooth disturbance between time stamps to characterize the evolution of networks for incrementally update.
- MMoE: Robust Spoiler Detection with Multi-modal Information and Domain-aware Mixture-of-Experts5
A multi-modal network that utilizes information from multiple modalities to facilitate robust spoiler detection and adopts Mixture-of-Experts to enhance domain generalization, and achieves state-of-the-art performance on two widely-used spoiler detection datasets.
- Sparse Relational Topical Coding on multi-modal data5
This paper presents a non-probabilistic formulation of Relational Topic Model (RTM), i.e., Sparse Relational Multi-Modal Topical Coding (SRMMTC), to model both multi-modal documents and the corresponding link information, and proposes a generalized version of SRMMTC to find pairwise interactions amongst topics.
- 4
- 4
- 4
- Towards Explanation for Unsupervised Graph-Level Representation Learning4
The Information Bottleneck principle is advanced and the connection between graph representations and explanatory subgraphs on the label space is theoretically analyzed, which reveals that the expressiveness and robustness of representations benefit the fidelity of explanatory subgraphs.
- Memory transformation networks for weakly supervised visual classification4
A Memory Transformation Network (MTN) is proposed to convert external knowledge, by involving embedded and concentrated memories, so as to leverage it feasibly for video classification with weak supervision, and outperforms the state-of-the-arts on visual classification with limited labeled exemplars.
- How Unlabeled Web Videos Help Complex Event Detection?4
A new robust dictionary learning framework for complex event detection, able to handle both labeled and easy-to-get unlabeled web videos by sharing the same dictionary is proposed, which shows strong robustness against the substantial noisy and outlier videos from open source.
- 4
- 3
- From Ideal to Real: Unified and Data-Efficient Dense Prediction for Real-World Scenarios3
This work introduces DenseWorld, a benchmark spanning a broad set of 25 dense prediction tasks that correspond to urgent real-world applications, featuring unified evaluation across tasks, and proposes DenseDiT, which exploits generative models'visual priors to perform diverse real-world dense prediction tasks through a unified strategy.
- 3
- 3
- OSPC: OCR-Assisted VLM for Zero-Shot Harmful Meme Detection3
This technical report summarises the fourth place solution of the Online Safety Prize Challenge, which enhances a visual language model (VLM) by integrating external Optical Character Recognition (OCR) tools, employs prompt engineering and manual resource management to optimize performance within limited resource constraints.
- 3
- 3
- 3
- The Facade of Truth: Uncovering and Mitigating LLM Susceptibility to Deceptive Evidence2
This work introduces MisBelief, a framework that generates misleading evidence via collaborative, multi-round interactions among multi-role LLMs, and proposes Deceptive Intent Shielding (DIS), a governance mechanism that provides an early warning signal by inferring the deceptive intent behind evidence.
- 2
- 2
- 2
- Unveiling the Hidden: Movie Genre and User Bias in Spoiler Detection2
A new spoiler detection framework called GUSD (Genre-aware and User-specific Spoiler Detection), which incorporates genre-specific data and user behavior bias, and achieves state-of-the-art results.
- 2
- 2
- Unsupervised Hierarchical Feature Selection on Networked Data2
A novel unsupervised feature selection framework (HNFS) that can simultaneously learn the implicit hierarchical structure among the nodes and embed the hierarchical structure into the feature selection process is proposed.
- Cross-Graph Representation Learning for Unsupervised Graph Alignment2
An unsupervised cross-graph representation learning framework to jointly learn the node representations of different graphs in a unified deep model based on both attribute and structure reconstruction and introduces a discriminator to better align the learned representations for different graphs via adversarial training.
- Deploying external bandwidth guaranteed media server clusters for real-time live streaming in media cloud2
The evaluation results show that the proposed scheme guarantees the external bandwidth requirement of a real-time live streaming application, at the same time, greatly reduces the intra-bandwidth utilization of a media cloud with different DCN structures.
- Heterogeneous Information Network Hashing for Fast Nearest Neighbor Search2
This paper proposes a novel heterogeneous information network hashing algorithm based on collective matrix factorization that can lead to faster NNS and requires lower memory usage than several state-of-the-art network embedding methods while showing comparable performance in typical learning tasks on HINs, including link prediction and cross-type node similarity search.
- 2
- 1
- Bot Meets Shortcut: How Can LLMs Aid in Handling Unknown Invariance OOD Scenarios?1
An in-depth study to assess how detectors are influenced by potential shortcuts based on textual features, which are most susceptible to manipulation by social bots, and proposes mitigation strategies based on large language models, leveraging counterfactual data augmentation.
- The Deliberative Illusion: Diagnosing Factual Attrition and Stance Homogenization in Multi-Agent LLM Deliberation1
The deliberative illusion is identified, and DelibTrace, a framework that decomposes each issue into atomic facts, labels issue-critical ones, distributes them across agents, and tracks their survival across discussion rounds is introduced, to measure which facts, uncertainties, and legitimate disagreements survive interaction.
- GuessBench: Sensemaking Multimodal Creativity in the Wild1
Analysis reveals that VLM performance in creativity sensemaking correlates with the frequency of the concept in training data, while the accuracy drops sharply for concepts in underrepresented cultural contexts and low-resource languages.
- 1
- Why Settle for One? Text-to-ImageSet Generation and Evaluation1
A training-free framework that maximally leverages pretrained Diffusion Transformers'in-context capabilities to harmonize visual elements to satisfy both image-level prompt alignment and set-level visual consistency, and significantly outperforms current generalized and even specialized approaches.
- 1
- 1
- Multi-Modal Dataset Distillation in the Wild1
Multi-modal dataset Distillation in the Wild is proposed, the first framework to distill noisy multi-modal datasets into compact clean ones for effective and efficient model training and introduces learnable fine-grained correspondences during distillation and adaptively optimizes distilled data to emphasize correspondence-discriminative regions, thereby enhancing distilled data's information density and efficacy.
- 1
- 1
- 1
- Disentangled Generation with Information Bottleneck for Few-Shot Learning1
A novel Information Bottleneck (IB) based Disentangled Generation Framework for FSL, termed as DisGenIB, that can simultaneously guarantee the discrimination and diversity of generated samples and can effectively utilize priors to further facilitate disentanglement.
- 1
- 1
- 1
- 1
- 1
- 1
- –
- –
- –
- –
- Correspondence Coverage Matters for Multi-Modal Dataset Distillation–
This work proposes a novel multi-modal DD framework to systematically Promote Correspondence coverage, i.e., ProCo, and develops a correspondence consistency metric based on cross-modal retrieval distributions to cluster correspondence patterns, enabling ProCo to initialize distilled data with representative patterns while regularizing optimization.
- Prompt Learning for Textual Heterogeneous Information Networks–
This work is among the first to introduce and exploit the idea of prompt learning to align HIN features and textual features, and is composed of a text encoder and a graph encoder and uses contrastive learning to align and fuse the graph–text pair.
- Dynamic graph representation learning with disentangled information bottleneck–
This work proposes to learn macro-disentangled dynamic graph representations based on the Information Bottleneck theory, leading to a novel dynamic graph representation learning method, Disentangled Dynamic Graph Information Bottleneck (DDGIB), and theoretically proves the sufficiency and macro disentanglement of DDGIB.
- –
- How Brittle is Agent Safety? Rethinking Agent Risk under Intent Concealment and Task Complexity–
OASIS (Orthogonal Agent Safety Inquiry Suite), a hierarchical benchmark with fine-grained annotations and a high-fidelity simulation sandbox, is introduced to provide a principled foundation for probing and strengthening agent safety in these overlooked dimensions.
- PTCL: Pseudo-Label Temporal Curriculum Learning for Label-Limited Dynamic Graph–
Experiments across real-world scenarios demonstrate PTCL's consistent superiority over other methods adapted to this task, and proposes a unified framework FLiD (Framework for Label-Limited Dynamic Node Classification), consisting of a complete preparation workflow, training pipeline, and evaluation standards, and supporting various models and datasets.
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-12. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.