IJCAI 2024 Papers — Page 3
International Joint Conference on Artificial Intelligence · 790 papers
Dirichlet-based Uncertainty Quantification for Personalized Federated Learning with Improved Posterior Networks
Nikita Kotelevskii (Mohamed bin Zayed University of Artificial Intelligence), Maxim Panov (Mohamed bin Zayed University of Artificial Intelligence)
Federated LearningFlow-based ModelImage
🎯 What it does: Proposes the FedPN framework, which dynamically switches between local and global models based on the confidence of the local and global models for input points, leveraging uncertainty quantification to improve prediction performance on out-of-distribution data.
Discriminative Feature Decoupling Enhancement for Speech Forgery Detection
Yijun Bei (Zhejiang University), Zunlei Feng (Zhejiang University)
ClassificationAnomaly DetectionConvolutional Neural NetworkGraph Neural NetworkContrastive LearningAudio
🎯 What it does: Propose a DEEM framework that decomposes speech into voice features and background noise features through exchange decoupling, further enhancing their discriminative features via time-domain and frequency-domain aggregation, and finally performing discrimination using a heterogeneous graph attention network.
Disentangling Domain and General Representations for Time Series Classification
Youmin Chen (Zhejiang University), Juren Li (Zhejiang University)
ClassificationDomain AdaptationRepresentation LearningConvolutional Neural NetworkTime Series
🎯 What it does: Proposes the CADT framework for unsupervised domain adaptation on time series, decoupling domain-invariant and domain-specific features.
Distribution-Independent Cell Type Identification for Single-Cell RNA-seq Data
Yuyao Zhai (Peking University), Minghua Deng (Peking University)
ClassificationAuto EncoderContrastive LearningBiomedical Data
🎯 What it does: This study proposes a distribution-agnostic, long-tailed open-domain single-cell RNA-seq cell type identification framework called scDET, addressing the problem of simultaneously performing classification and clustering in data with scarce annotations and containing unknown cell types.
Diversification of Adaptive Policy for Effective Offline Reinforcement Learning
Yunseon Choi (KAIST AI), Kee-Eung Kim (KAIST AI)
Recurrent Neural NetworkReinforcement LearningBenchmark
🎯 What it does: This paper proposes a model-based offline reinforcement learning method called MoDAP, which trains iterable multi-model dynamics on offline data and enables the policy to adaptively select actions through a recurrent network, addressing decision-making challenges beyond the data support.
Diversifying Training Pool Predictability for Zero-shot Coordination: A Theory of Mind Approach
Dung Nguyen (Deakin University), Truyen Tran (Deakin University)
Reinforcement LearningSequential
🎯 What it does: In zero-shot coordination, a diverse training pool is constructed by assessing and maintaining the behavioral predictability (LoBP) of collaboration partners through a Theory of Mind (ToM) observer, enhancing the adaptive agent's ability to coordinate with unknown partners.
Domain Adaptive and Fine-grained Anomaly Detection for Single-cell Sequencing Data and Beyond
Kaichen Xu (Zhongnan University of Economics and Law), Xiaobo Sun (Zhongnan University of Economics and Law)
Domain AdaptationAnomaly DetectionGenerative Adversarial NetworkBiomedical Data
🎯 What it does: Propose a generative framework named ACSleuth that unifies anomaly cell detection, domain adaptation, and fine-grained anomaly cell classification in single-cell sequence data;
Domain-Hierarchy Adaptation via Chain of Iterative Reasoning for Few-shot Hierarchical Text Classification
Ke Ji (Southeast University), Ziyu Shang (Southeast University)
ClassificationDomain AdaptationTransformerPrompt EngineeringTextChain-of-Thought
🎯 What it does: Propose the HierICRF framework, combining hierarchical reasoning chains with hierarchical iterative CRF, to accomplish hierarchical text classification tasks under extremely few samples.
DTS-TPT: Dual Temporal-Sync Test-time Prompt Tuning for Zero-shot Activity Recognition
Rui Yan (Nanjing University), Tieniu Tan (Nanjing University)
RecognitionLarge Language ModelPrompt EngineeringVision Language ModelVideo
🎯 What it does: This paper proposes the Dual-Temporal Synchronization Test-Time Prompt Tuning framework (DTS-TPT) for zero-shot video action recognition.
Dual Calibration-based Personalised Federated Learning
Xiaoli Tang (Nanyang Technological University), Xiaoxiao Li (University of British Columbia)
Federated LearningImage
🎯 What it does: Propose the DC-PFL method, which achieves collaborative training of personalized feature extractors and shared classifiers through dual calibration in federated learning.
Dual Contrastive Graph-Level Clustering with Multiple Cluster Perspectives Alignment
Jinyu Cai (National University of Singapore), Wenzhong Guo (Fuzhou University)
Representation LearningGraph Neural NetworkContrastive LearningGraph
🎯 What it does: Proposed a dual-contrast graph hierarchical clustering method DCGLC, achieving unsupervised graph layer clustering through graph contrastive learning and multi-view clustering alignment.
Dual Enhancement in ODI Super-Resolution: Adapting Convolution and Upsampling to Projection Distortion
Xiang Ji (Beijing University of Posts and Telecommunications), Gabriel-Miro Muntean (Dublin City University)
Super ResolutionConvolutional Neural NetworkImage
🎯 What it does: Developed a Dual-Enhanced Super-Resolution Network (ODA-SRN) for panoramic image projection distortion problems, achieving efficient panoramic image super-resolution through improved convolution and upsampling.
Dual Expert Distillation Network for Generalized Zero-Shot Learning
Zhijie Rao (Hong Kong Polytechnic University), Xiaofeng Cao (Jilin University)
ClassificationKnowledge DistillationRepresentation LearningConvolutional Neural NetworkMixture of ExpertsImageBenchmark
🎯 What it does: Propose the Dual Expert Distillation Network (DEDN), which learns visual-attribute associations through a coarse-grained expert (cExp) and a fine-grained expert (fExp), introduces a Dual Attention Network (DAN) to integrate spatial and channel attention, and proposes a Margin-Aware Loss (MAL) to balance the confidence of seen and unseen classes.
Dual Semantic Fusion Hashing for Multi-Label Cross-Modal Retrieval
Kaiming Liu (Sichuan University), Yuan Sun (Sichuan University)
RetrievalImageTextMultimodalityBenchmark
🎯 What it does: Propose a Dual Semantic Fusion Hashing (DSFH) framework for multi-label cross-modal retrieval. By learning modality-specific representations and consensus hash codes, and fusing coarse-grained (instance-level) and fine-grained (cluster-level) semantic relationships to generate more discriminative binary codes.
DVPE: Divided View Position Embedding for Multi-View 3D Object Detection
Jiasen Wang (Shanghai University), Yang Zhou (Shanghai University)
Object DetectionAutonomous DrivingTransformerImageTime Series
🎯 What it does: Proposes a divided-view position embedding (DVPE) framework for multi-view 3D object detection, integrating separable view visibility attention, temporal modeling of 2D RoI priors, and a one-to-many assignment training strategy.
DWLR: Domain Adaptation under Label Shift for Wearable Sensor
Juren Li (Zhejiang University), Lujia Pan (Huawei Noah's Ark Lab)
Domain AdaptationConvolutional Neural NetworkGenerative Adversarial NetworkTime SeriesBiomedical Data
🎯 What it does: Proposed the DWLR method to address the problem of simultaneous feature drift and label drift in wearable sensors under unlabeled target domains.
Dynamic against Dynamic: An Open-Set Self-Learning Framework
Haifeng Yang (Nanjing University of Aeronautics and Astronautics), Songcan Chen (Nanjing University of Aeronautics and Astronautics)
ClassificationRecognitionImageBenchmark
🎯 What it does: Propose an open-set self-learning (OSSL) framework based on the dynamic-to-dynamic idea, designed to dynamically adapt to unknown class samples and enhance recognition performance in open and dynamic scenarios.
Dynamic Brightness Adaptation for Robust Multi-modal Image Fusion
Yiming Sun (Tianjin University), Qinghua Hu (Tianjin University)
RestorationConvolutional Neural NetworkMultimodality
🎯 What it does: Propose a multi-modal image fusion framework BA-Fusion based on brightness-adaptive gating, which can achieve robust infrared-visible image fusion under variable brightness conditions.
Dynamic Many-Objective Molecular Optimization: Unfolding Complexity with Objective Decomposition and Progressive Optimization
Dong-Hee Shin (Korea University), Tae-Eui Kam (Korea University)
OptimizationDrug DiscoveryTransformerReinforcement LearningBiomedical DataBenchmark
🎯 What it does: Propose a DyMol method that addresses dynamic multi-objective molecular optimization problems using the ideas of decomposition and progressive optimization.
Dynamic Weighted Graph Fusion for Deep Multi-View Clustering
Yazhou Ren (University of Electronic Science and Technology of China), Lifang He (Lehigh University)
Graph Neural NetworkAuto EncoderMultimodality
🎯 What it does: This paper proposes a dynamic weighted graph fusion deep multi-view clustering framework called DFMVC. It first learns the embedded features of each view through an autoencoder, then constructs a potential graph and concatenates the features of all views to obtain global features. A graph convolution module is used to perform self-supervised learning on the fused graph, ultimately achieving clustering results.
Dynamically Anchored Prompting for Task-Imbalanced Continual Learning
Chenxing Hong (Xiamen University), Hanzi Wang (Xiamen University)
ClassificationTransformerPrompt EngineeringImage
🎯 What it does: Proposed a dynamic anchoring prompt (DAP) method to address the task imbalance problem in continual learning;
Dynamicity-aware Social Bot Detection with Dynamic Graph Transformers
Buyun He (University of Science and Technology of China), Pengyuan Zhou (Aarhus university)
ClassificationAnomaly DetectionGraph Neural NetworkTransformerGraph
🎯 What it does: This paper proposes a new social bot detection framework called BotDGT, focusing on the dynamics of social networks, combining a structural module and a temporal module to capture the context of historical interaction graphs and evolutionary behaviors;
EAB-FL: Exacerbating Algorithmic Bias through Model Poisoning Attacks in Federated Learning
Syed Irfan Ali Meerza (University of Tennessee, Knoxville), Jian Liu (University of Tennessee, Knoxville)
Federated LearningAdversarial AttackImageTabular
🎯 What it does: A technique leveraging model redundancy space for model poisoning attacks is proposed within the federated learning framework to significantly exacerbate group unfairness while maintaining overall model accuracy.
EAT: Self-Supervised Pre-Training with Efficient Audio Transformer
Wenxi Chen (Shanghai Jiao Tong University), Xie Chen (Shanghai Jiao Tong University)
ClassificationComputational EfficiencyRepresentation LearningConvolutional Neural NetworkTransformerAudio
🎯 What it does: Propose the EAT model, which employs a self-supervised trained audio Transformer, combined with inverse block multiple masking and Utterance-Frame Objective to enhance pre-training efficiency and performance.
EC-SNN: Splitting Deep Spiking Neural Networks for Edge Devices
Di Yu (Zhejiang University), Shuiguang Deng (Zhejiang University)
Computational EfficiencyKnowledge DistillationSpiking Neural NetworkImageAudio
🎯 What it does: Split large deep SNNs into multiple sub-models and parallelize their execution on multiple edge devices, further compressing the sub-models using channel pruning.
ECR-Chain: Advancing Generative Language Models to Better Emotion-Cause Reasoners through Reasoning Chains
Zhaopei Huang (Renmin University of China), Qin Jin (Renmin University of China)
ClassificationRecognitionLarge Language ModelSupervised Fine-TuningTextChain-of-Thought
🎯 What it does: Propose an ECR-Chain based on cognitive appraisal theory, which follows a 'Theme→Response→Assessment→Stimulus' reasoning chain. This approach guides large models like ChatGPT to perform emotion causality reasoning through few-shot prompting. Additionally, large models are utilized to automatically construct the ECR-Chain dataset for multi-task training of small models, significantly enhancing emotion causality identification and explainable reasoning capabilities.
EFEVD: Enhanced Feature Extraction for Smart Contract Vulnerability Detection
Chi Jiang (University of Electronic Science and Technology of China), Yin Zhang (University of Electronic Science and Technology of China)
ClassificationAnomaly DetectionConvolutional Neural NetworkGraph Neural NetworkTransformerTextGraphFinance Related
🎯 What it does: This paper proposes an enhanced feature extraction method (EFEVD) that improves smart contract vulnerability detection through community features and cross-task shared features.
Effective Approach to LTLf Best-Effort Synthesis in Multi-Tier Environments
Benjamin Aminof (Universit' a degli Studi di Roma 'La Sapienza'), Sasha Rubin (University of Sydney)
OptimizationComputational EfficiencyBenchmark
🎯 What it does: This paper studies best-effort synthesis using LTLf in multi-tier environments and proposes an algorithm that completes in linear time.
Efficiency Calibration of Implicit Regularization in Deep Networks via Self-paced Curriculum-Driven Singular Value Selection
Zhe Li (Nanjing University of Science and Technology), Lei Luo (Nanjing University of Science and Technology)
Recommendation SystemOptimizationComputational EfficiencyTabular
🎯 What it does: Propose an adaptive weighted nuclear norm (SPWNN) to achieve matrix recovery through a linear neural network, combining explicit and implicit regularization to adjust low-rank bias.
Efficient and Stable Offline-to-online Reinforcement Learning via Continual Policy Revitalization
Rui Kong (Nanjing University), Ming Li (Nanjing University)
Reinforcement LearningBenchmark
🎯 What it does: Propose a Continuous Policy Revival (CPR) method in offline-to-online reinforcement learning, achieving stable and efficient online fine-tuning through periodic policy reset, retaining historical policy sets, and using adaptive policy constraints.
Efficient Correlated Subgraph Searches for AI-powered Drug Discovery
Hiroaki Shiokawa (University of Tsukuba), Shohei Matsugu (University of Tsukuba)
Drug DiscoveryGraph
🎯 What it does: Propose the Corgi framework, which efficiently accelerates the search for relevant subgraphs (CSS) in drug databases by utilizing a summary graph view and multi-view Top-k search.
Efficient Cost-Minimization Schemes for Electrical Energy Demand Satisfaction by Prosumers in Microgrids with Battery Storage Capabilities
Laura Codazzi (Hamburg University of Technology), Matthias Mnich (Hamburg University of Technology)
OptimizationTabularTime Series
🎯 What it does: This paper studies the problem of minimizing energy costs for prosumers in microgrids over a fixed time period through battery storage and buy/sell transactions with other prosumers, and provides multiple integer linear programming models as well as efficient combinatorial and dynamic programming algorithms;
Efficient Event Stream Super-Resolution with Recursive Multi-Branch Fusion
Quanmin Liang (Sun Yat-Sen University), Yonghong Tian (Peng Cheng Laboratory)
Super ResolutionRecurrent Neural Network
🎯 What it does: Propose a recursive multi-branch information fusion network RMFNET, which extracts and fuses information using positive/negative event branches and an event frame branch to achieve efficient event stream super-resolution.
Efficient Federated Multi-View Clustering with Integrated Matrix Factorization and K-Means
Wei Feng (Xi'an Jiaotong University), Quanxue Gao (Xi'an Jiaotong University)
Federated LearningSafty and PrivacyMultimodality
🎯 What it does: Propose a federated multi-view clustering method FMVC-IMK, which integrates matrix decomposition with K-means into a single-step optimization and incorporates graph regularization to preserve local geometric structures.
Efficient Multi-view Unsupervised Feature Selection with Adaptive Structure Learning and Inference
Chenglong Zhang (Hangzhou Normal University), Weiguo Sheng (Hangzhou Normal University)
Computational EfficiencyRepresentation LearningImageMultimodality
🎯 What it does: Proposes an efficient multi-view unsupervised feature selection method called EMUFS, which selects important features across multiple views by leveraging anchor-based bilateral graphs, clustering information inference, and adaptive structure learning.
Efficient Offline Meta-Reinforcement Learning via Robust Task Representations and Adaptive Policy Generation
Zhengwei Li (University of Chinese Academy of Science), Zhiyong Liu (University of Chinese Academy of Science)
Representation LearningMeta LearningReinforcement LearningAuto EncoderContrastive LearningBenchmark
🎯 What it does: Proposes a novel offline meta-reinforcement learning framework R2PGO, which achieves zero-shot adaptation and efficient learning by combining robust task representation with adaptive strategy generation.
Efficient Screen Content Image Compression via Superpixel-based Content Aggregation and Dynamic Feature Fusion
Sheng Shen (Tianjin University), Jingyu Yang (Tianjin University)
CompressionTransformerAuto EncoderImage
🎯 What it does: This paper proposes a learning-based screen content image compression network based on superpixel aggregation and dynamic feature fusion, specifically designed for efficient compression of screen content images (SCI);
Efficient Tuning and Inference for Large Language Models on Textual Graphs
Yun Zhu (Zhejiang University), Siliang Tang (Zhejiang University)
ClassificationComputational EfficiencyGraph Neural NetworkTransformerLarge Language ModelGraph
🎯 What it does: Proposes ENGINE, a parameter and memory-efficient fine-tuning method for large language models (LLMs) on text graphs, which employs a frozen LLM with a lightweight G-Ladder side structure, and further enhances training and inference speed through caching and dynamic early stopping.
ELF-UA: Efficient Label-Free User Adaptation in Gaze Estimation
Yong Wu (Tongji University), Guang Chen (Tongji University)
Pose EstimationDomain AdaptationMeta LearningConvolutional Neural NetworkContrastive LearningImage
🎯 What it does: Propose an efficient, unlabeled user-adaptive gaze estimation method that personalizes the model using a small number of unlabeled user images.
Eliminating the Cross-Domain Misalignment in Text-guided Image Inpainting
Muqi Huang (Wuhan University), Lefei Zhang (Wuhan University)
RestorationGenerationVision Language ModelDiffusion modelImageText
🎯 What it does: Propose a structure-aware training scheme (SAIL) and an asymmetric cross-domain attention mechanism (ACDA) to improve text-guided image inpainting, making the generated results more structurally and texturally consistent with the original image and text description.
EMOTE: An Explainable Architecture for Modelling the Other through Empathy
Manisha Senadeera (Deakin University), Santu Rana (Deakin University)
Explainability and InterpretabilityReinforcement Learning
🎯 What it does: This paper proposes EMOTE (Explainable Architecture for Modelling the Other through Empathy), which models the action value and reward function of independent agents by learning the Q-function of the agent itself, and achieves online inverse reinforcement learning through a two-stage network that generates explainable empathy states;
Empirical Analysis of Dialogue Relation Extraction with Large Language Models
Guozheng Li (Southeast University), Yikai Guo (Beijing Institute of Computer Technology and Application)
TransformerLarge Language ModelSupervised Fine-TuningPrompt EngineeringTextMultimodality
🎯 What it does: Explored the application of large language models in dialogue relation extraction, including prompt-based extraction for ChatGPT and LoRA fine-tuning for open-source LLMs.
Enabling Mixed Effects Neural Networks for Diverse, Clustered Data Using Monte Carlo Methods
Andrej Tschalzev (University of Mannheim), Heiner Stuckenschmidt (University of Mannheim)
ClassificationExplainability and InterpretabilityData-Centric LearningTabularBenchmark
🎯 What it does: Proposed and implemented a mixed-effect neural network training framework named MC-GMENN, capable of simultaneously handling multi-class and multi-cluster features, and achieving precise Bayesian estimation of random effects through Monte Carlo EM and NUTS sampling.
Enabling Sustainable Freight Forwarding Network via Collaborative Games
Pang-Jin Tan (Singapore Management University), Richard Chen (Coupang)
OptimizationGraph
🎯 What it does: This paper proposes the 'Local Collaboration Game' (LCG) model to describe cooperative games in logistics networks where agents can only collaborate with their neighbors, and designs an efficient algorithm called FS-LCG to compute the Shapley value for this model; subsequently, it applies this model to the Forward Agent Collaboration Game (FFCG), achieving fair cost allocation;
Encoding Auxiliary Information to Restore Compressed Point Cloud Geometry
Gexin Liu (Hangzhou Normal University), Zhan Ma (Nanjing University)
RestorationCompressionConvolutional Neural NetworkPoint Cloud
🎯 What it does: Proposes an AuxGR framework that utilizes an auxiliary bitstream to recover the geometry of G-PCC compressed point clouds.
Endogenous Energy Reactive Modules Games: Modelling Side Payments among Resource-Bounded Agents
Julian Gutierrez (Monash University), Michael Wooldridge (University of Oxford)
🎯 What it does: Proposes the Energy Reaction Module Game (ERMG) and its endogenous version (EERMG) for modeling multi-agent systems with energy constraints, and investigates the rational verification problem within this framework;
Enhanced DouDiZhu Card Game Strategy Using Oracle Guiding and Adaptive Deep Monte Carlo Method
Qian Luo (Universiti Sains Malaysia), Tianqiao Zhang (Guilin University of Electronic Technology)
Reinforcement LearningTabular
🎯 What it does: Propose a DouDiZhu AI named OADMCDou that integrates Oracle guidance with adaptive deep Monte Carlo (ADMC), achieving performance surpassing existing DouZero through self-play training.
Enhancing Boundary Segmentation for Topological Accuracy with Skeleton-based Methods
Chuni Liu (University of Science and Technology Beijing), Ke Xu (University of Science and Technology Beijing)
SegmentationConvolutional Neural NetworkImage
🎯 What it does: This paper proposes a Skea-Topo Aware loss function for boundary segmentation tasks, significantly improving the topological consistency of segmentation results.
Enhancing Controlled Query Evaluation through Epistemic Policies
Gianluca Cima (Sapienza University of Rome), Domenico Fabio Savo (University of Bergamo)
Safty and Privacy
🎯 What it does: Proposes using epistemic dependencies (ED) in DL-Lite R ontologies to express information protection policies, and computes secure query answers through optimal CQ-censor;
Enhancing Cooperation through Selective Interaction and Long-term Experiences in Multi-Agent Reinforcement Learning
Tianyu Ren (University of Manchester), Xiao-Jun Zeng (University of Manchester)
Reinforcement LearningGraph
🎯 What it does: In an iterated prisoner's dilemma game on a 2D grid, researchers designed a multi-agent reinforcement learning framework enabling agents to simultaneously learn cooperation/betrayal strategies and neighbor selection strategies under the guidance of long-term experience.
Enhancing Cross-modal Completion and Alignment for Unsupervised Incomplete Text-to-Image Person Retrieval
Tiantian Gong (Nanjing University of Aeronautics and Astronautics), Liyan Zhang (Nanjing University of Aeronautics and Astronautics)
RetrievalTransformerContrastive LearningImageTextMultimodality
🎯 What it does: Proposes the ECCA framework for unsupervised and incomplete text-image person retrieval;
Enhancing Cross-Modal Retrieval via Visual-Textual Prompt Hashing
Bingzhi Chen (Beijing Institute of Technology), Zheng Zhang (Harbin Institute of Technology)
RetrievalTransformerLarge Language ModelPrompt EngineeringContrastive LearningImageTextMultimodality
🎯 What it does: Propose a Visual-Text Prompt Hashing (VTPH) framework that enhances cross-modal retrieval performance by leveraging visual-text prompt learning and adaptive contrastive learning.
Enhancing Dual-Target Cross-Domain Recommendation with Federated Privacy-Preserving Learning
Zhenghong Lin (Fuzhou University), Yanchao Tan (Fuzhou University)
Recommendation SystemFederated LearningSafty and PrivacyGraph Neural NetworkTabular
🎯 What it does: Proposed the P2DTR framework, which realizes a federated privacy-preserving recommendation system in dual-target cross-domain recommendation (DTCDR), addressing the privacy leakage issues caused by traditional methods sharing user sensitive data.
Enhancing Fine-Grained Urban Flow Inference via Incremental Neural Operator
Qiang Gao (Southwestern University of Finance and Economics), Xueqin Chen (Delft University of Technology)
OptimizationSafty and PrivacyConvolutional Neural NetworkTime SeriesSequential
🎯 What it does: Proposes a fine-grained traffic inference framework UNOI (Urban Neural Operator with Incremental Learning) based on neural operators and incremental learning, which can predict fine-grained traffic with only coarse-grained inputs, and addresses catastrophic forgetting and privacy leakage issues through two incremental strategies.
Enhancing Length Generalization for Attention Based Knowledge Tracing Models with Linear Biases
Xueyi Li (Jinan University), Jian Weng (Jinan University)
TransformerTabularSequential
🎯 What it does: Propose a knowledge tracing model named stableKT, which can maintain high accuracy on long sequences after being trained on short sequences;
Enhancing Multimodal Knowledge Graph Representation Learning through Triple Contrastive Learning
Yuxing Lu (Peking University), Jinzhuo Wang (Peking University)
Representation LearningGraph Neural NetworkTransformerLarge Language ModelContrastive LearningMultimodalityBiomedical Data
🎯 What it does: Proposed the KG-MRI method, which utilizes triple contrastive learning and two-stage training to fuse multimodal knowledge graph entity representations.
Enhancing Scalability of Metric Differential Privacy via Secret Dataset Partitioning and Benders Decomposition
Chenxi Qiu (University of North Texas)
OptimizationSafty and PrivacyComputational EfficiencyTextGraphTabular
🎯 What it does: Propose a computational framework based on secret dataset partitioning and Benders decomposition, significantly enhancing the scalability of Metric Differential Privacy (mDP) linear programming.
ENOTO: Improving Offline-to-Online Reinforcement Learning with Q-Ensembles
Kai Zhao (Tianjin University), Zhaopeng Meng (Tianjin University)
Reinforcement LearningMixture of ExpertsTabularTime SeriesBenchmark
🎯 What it does: Propose the ENOTO (Ensemble-based Offline-to-Online) framework, which utilizes Q-ensembles to achieve a seamless transition between offline pre-training and online fine-tuning. By appropriately relaxing pessimism and incorporating ensemble-based exploration mechanisms, the stability, learning efficiency, and final performance of offline-to-online reinforcement learning are significantly improved.
EPIC: Graph Augmentation with Edit Path Interpolation via Learnable Cost
Jaeseung Heo (POSTECH), Dongwoo Kim (POSTECH)
ClassificationData SynthesisGraph Neural NetworkGraph
🎯 What it does: This paper proposes a graph edit path interpolation data augmentation method called EPIC based on learnable edit costs;
Epistemic Logic Programs: Non-Ground and Counting Complexity
Thomas Eiter (TU Wien), Stefan Woltran (TU Wien)
🎯 What it does: This paper studies the computational complexity of non-ground Epistemic Logic Programs (ELP) in qualitative and quantitative reasoning, and provides a complete complexity landscape.
Equilibria in Two-Stage Facility Location with Atomic Clients
Simon Krogmann (University of Potsdam), Marnix C. Vos (University of Twente)
Optimization
🎯 What it does: In a two-stage competitive facility location model considering atomic (indivisible) customers, it is proven that a subgame perfect equilibrium (SPE) exists when customer weights are equal, but SPE may not exist when weights differ, and determining an approximate SPE is NP-hard; meanwhile, the price upper bound for SPE is 2.
Error-aware Sampling in Adaptive Shells for Neural Surface Reconstruction
Qi Wang (Zhejiang University), Hujun Bao (Zhejiang University)
Neural Radiance FieldPoint CloudMesh
🎯 What it does: The paper proposes an error-aware sampling method based on learnable spatially varying kernel sizes, and achieves significant reduction in the number of sampling points while maintaining reconstruction accuracy through a dual cropping strategy to obtain an adaptive shell.
ESP-PCT: Enhanced VR Semantic Performance through Efficient Compression of Temporal and Spatial Redundancies in Point Cloud Transformers
Luoyu Mei (Southeast University), Wei Gong (University of Science and Technology of China)
RecognitionRecurrent Neural NetworkTransformerPoint CloudSequentialBenchmark
🎯 What it does: Developed a two-stage semantic recognition framework named ESP-PCT, leveraging sparse point clouds generated by millimeter-wave radar to achieve efficient target localization and focusing in VR scenes, thereby improving semantic recognition accuracy while significantly reducing computational and storage costs.
Estimating before Debiasing: A Bayesian Approach to Detaching Prior Bias in Federated Semi-Supervised Learning
Guogang Zhu (Beihang University), Hao Su (Beihang University)
Federated LearningImage
🎯 What it does: Propose FedDB, a Bayesian debiasing method for federated semi-supervised learning, which improves the model's generalization performance on imbalanced data by estimating and correcting label prior bias.
Estimating Conditional Average Treatment Effects via Sufficient Representation Learning
Pengfei Shi (Xiamen University), Yin Jin (Ant Group)
Representation LearningTabular
🎯 What it does: Propose a neural network framework called CrossNet to estimate the conditional average treatment effect (CATE), by learning sufficient representations that satisfy the no-confounding assumption, and simultaneously utilizing both self-group and cross-group data when training the prediction function.
Evaluation of Project Performance in Participatory Budgeting
Niclas Boehmer (Harvard University), Stanisław Szufa (AGH University)
OptimizationComputational EfficiencyTabularFinance Related
🎯 What it does: The paper proposes a new method to evaluate the performance of failed projects in participatory budgeting, quantifying the cost reductions, additional support required, or competitor support cuts needed for them to win.
EvaNet: Elevation-Guided Flood Extent Mapping on Earth Imagery
Mirza Tanzim Sami (University of Alabama at Birmingham), Yang Zhou (Auburn University)
SegmentationConvolutional Neural NetworkImage
🎯 What it does: Proposed EvaNet, a flood extent segmentation network guided by digital elevation models (DEM).
EVE: Efficient Zero-Shot Text-Based Video Editing With Depth Map Guidance and Temporal Consistency Constraints
Yutao Chen (Shandong University), Qingpei Guo (Ant Group)
GenerationDepth EstimationVision Language ModelDiffusion modelVideoTextBenchmark
🎯 What it does: Proposed EVE, a zero-copy text-driven video editing method based on pre-trained LDM, enhancing temporal consistency through depth map guidance and frame-aligned attention.
Evolutionary Generalized Zero-Shot Learning
Dubing Chen (Nanjing University of Science and Technology), Haofeng Zhang (Nanjing University of Science and Technology)
ClassificationDomain AdaptationKnowledge DistillationSupervised Fine-TuningImageBenchmark
🎯 What it does: Propose an evolutionary generalized zero-shot learning (EGZSL) framework that enables the model to continuously adapt and improve performance on unlabeled test streams.
Exactly Solving Minimum Dominating Set and Its Generalization
Ziliang Xiong (University of Electronic Science and Technology of China), Mingyu Xiao (University of Electronic Science and Technology of China)
OptimizationGraphBenchmark
🎯 What it does: Proposed two exact algorithms for solving the minimum dominating set problem based on branch-and-bound, using combinatorial lower bounds and LP relaxation lower bounds respectively.
Expected Work Search: Combining Win Rate and Proof Size Estimation
Owen Randall (University of Alberta), Ryan Hayward (University of Alberta)
OptimizationReinforcement Learning
🎯 What it does: Proposed a novel search algorithm called Expected Work Search (EWS), which predicts and minimizes the computational effort required to solve positions by simultaneously estimating win rates and proof sizes.
Explaining Arguments’ Strength: Unveiling the Role of Attacks and Supports
Xiang Yin (Imperial College London), Francesca Toni (Imperial College London)
Explainability and InterpretabilityComputational EfficiencyGraph
🎯 What it does: This paper proposes the Relation Attribution Explanation (RAE) framework, which utilizes Shapley values to provide fine-grained explanations for attack and support relationships in the Quantitative Bipolar Argumentation Framework (QBAF), and presents an efficient probabilistic approximation algorithm to compute RAE.
Exploiting Conjugate Label Information for Multi-Instance Partial-Label Learning
Wei Tang (Southeast University), Min-Ling Zhang (Southeast University)
ClassificationImageBenchmark
🎯 What it does: Propose the ELIMIPL algorithm, which improves the disambiguation performance in multi-instance partial label learning by leveraging conjugate information between candidate and non-candidate labels, as well as the sparsity of true labels.
Exploiting Multi-Label Correlation in Label Distribution Learning
Zhiqiang Kou (Southeast University), Xin Geng (Southeast University)
OptimizationRepresentation LearningImageTabularBiomedical Data
🎯 What it does: An auxiliary multi-label learning (MLL) process is proposed in label distribution learning (LDL), leveraging low-rank label correlations on the auxiliary process to enhance LDL's predictive performance; label distributions are converted into multi-labels via threshold or top-k methods, followed by jointly optimizing the mapping between label distributions and multi-labels, forming two methods: TLRLDL and TKLRLDL.
Explore Internal and External Similarity for Single Image Deraining with Graph Neural Networks
Cong Wang (Shenzhen Campus, Sun Yat-sen University), Jie Mu (Dongbei University of Finance and Economics)
RestorationConvolutional Neural NetworkGraph Neural NetworkImage
🎯 What it does: Proposed a multi-scale graph neural network (MSGNN) that utilizes internal multi-scale self-similarity and external example similarity to aggregate neighboring image patches via graph neural networks for single-image deraining.
Exploring Cross-Domain Few-Shot Classification via Frequency-Aware Prompting
Tiange Zhang (Ocean University of China), Junyu Dong (Ocean University of China)
ClassificationDomain AdaptationMeta LearningConvolutional Neural NetworkPrompt EngineeringImageBiomedical Data
🎯 What it does: Studied cross-domain few-shot classification, proposing Frequency-Aware Prompting and mutual attention mechanisms to enhance the cross-domain generalization ability of meta-learning models.
Exploring Learngene via Stage-wise Weight Sharing for Initializing Variable-sized Models
Shi-Yu Xia (Southeast University), Xin Geng (Southeast University)
Computational EfficiencyKnowledge DistillationTransformerSupervised Fine-TuningImage
🎯 What it does: Designed and implemented a phased weight sharing (SWS) framework to learn a compact learngene, and extended it to multi-scale Transformers to adapt to model deployment under different resource constraints.
Exploring the Inefficiency of Heavy Ball as Momentum Parameter Approaches 1
Xiaoge Deng (National University of Defense Technology), Xicheng Lu (National University of Defense Technology)
OptimizationImage
🎯 What it does: Investigated the inefficiency of the Heavy Ball method in stochastic gradient descent when the momentum parameter β approaches 1, and provided a theoretical explanation.
Exploring the Role of Node Diversity in Directed Graph Representation Learning
Jincheng Huang (University of Electronic Science and Technology of China), Xiaofeng Zhu (University of Electronic Science and Technology of China)
Representation LearningGraph Neural NetworkGraph
🎯 What it does: This paper proposes a new spatial directed graph neural network framework, NDDGNN, which adaptively assigns weights to incoming and outgoing information for each node by considering node diversity (neighbor diversity and degree diversity), thereby enhancing the representation learning effect on directed graphs.
Exploring Urban Semantics: A Multimodal Model for POI Semantic Annotation with Street View Images and Place Names
Dabin Zhang (Shandong University), Kai Zhao (Georgia State University)
ClassificationConvolutional Neural NetworkTransformerSupervised Fine-TuningVision Language ModelImageTextMultimodality
🎯 What it does: This paper proposes a multimodal model M3PA that fuses street view images with POI names and their neighbors' information to achieve semantic annotation of POIs.
Exponential Lower Bounds on the Double Oracle Algorithm in Zero-Sum Games
Brian Hu Zhang (Carnegie Mellon University), Tuomas Sandholm (Carnegie Mellon University)
OptimizationReinforcement Learning
🎯 What it does: This paper constructs several zero-sum game instances to prove that the double speculation algorithm may require exponentially many iterations to converge in partially observable stochastic games and tree games.
Expressiveness is Effectiveness: Self-supervised Fashion-aware CLIP for Video-to-Shop Retrieval
Likai Tian (Wuhan University), Zheng Wang (Wuhan University)
Pose EstimationRetrievalConvolutional Neural NetworkGraph Neural NetworkTransformerSupervised Fine-TuningVision Language ModelContrastive LearningVideoText
🎯 What it does: Propose a self-supervised Fashion-aware CLIP (SF-CLIP) framework, which evaluates the fashion expressiveness of video frames (occlusion, viewpoint, cropping) using pseudo labels to select high-expressive frames for video-store retrieval, and designs a dual-branch graph fusion module to integrate multi-frame information.
Exterior Penalty Policy Optimization with Penalty Metric Network under Constraints
Shiqing Gao (Shanghai Jiao Tong University), Chenghu Zhou (Shanghai Jiao Tong University)
Reinforcement LearningBenchmark
🎯 What it does: Propose an external penalty strategy optimization method called EPO, and design a Penalty Metric Network (PMN) to adaptively assign penalties for different levels of violation;
Extremal Separation Problems for Temporal Instance Queries
Jean Christoph Jung (TU Dortmund University), Michael Zakharyaschev (University of London)
OptimizationComputational EfficiencySequential
🎯 What it does: Studied the extremal separation problem for instance queries composed using the 'next' and 'finally' operators in linear temporal logic (LTL), determining the existence, counting, and construction complexity of the most specific/general separation queries across multiple query classes, and providing various polynomial-time algorithms.
Facility Location Problems with Capacity Constraints: Two Facilities and Beyond
Gennaro Auricchio (University of Bath), Jie Zhang (University of Bath)
Optimization
🎯 What it does: For the m-CFLP where facilities are collocated on a line with equal capacity and no excess capacity, two truthful computable mechanisms (PMM and PIPM) are proposed, and their limited approximation ratios are proven on social cost (SC) and maximum cost (MC). Additionally, for the 2-CFLP where two facilities can accommodate at least half of the agents, an extended internal gap (EIG) mechanism achieving strong group strategy-proofness is proposed, proven optimal on MC and nearly optimal on SC.
FactCHD: Benchmarking Fact-Conflicting Hallucination Detection
Xiang Chen (Zhejiang University), Huajun Chen (Alibaba Group)
Anomaly DetectionTransformerLarge Language ModelSupervised Fine-TuningTextBiomedical DataBenchmarkRetrieval-Augmented Generation
🎯 What it does: Constructed the FACTCHD benchmark to evaluate LLM's fact conflict hallucination detection, and proposed TRUTH-TRIANGULATOR for multi-source evidence triangulation.
Fair Distribution of Delivery Orders
Hadi Hosseini (Pennsylvania State University), Tomasz Wąs (University of Oxford)
OptimizationGraph
🎯 What it does: This paper studies a model for fairly allocating delivery tasks among multiple agents in a tree graph, exploring the compatibility between fairness (EF1, MMS) and efficiency (social optimality, Pareto optimality) and their algorithmic implementation.
FairGT: A Fairness-aware Graph Transformer
Renqiang Luo (Dalian University of Technology), Feng Xia (RMIT University)
ClassificationSafty and PrivacyTransformerGraph
🎯 What it does: Proposed a fairness-adaptive graph transformer called FairGT to address the bias issue in traditional Graph Transformers on sensitive features;
Fast and Continual Knowledge Graph Embedding via Incremental LoRA
Jiajun Liu (Southeast University), Yining Li (Southeast University)
Computational EfficiencyRepresentation LearningGraph
🎯 What it does: Propose the FastKGE framework, which utilizes an incremental low-rank adapter (IncLoRA) to enable efficient learning and storage for continuously growing knowledge graphs.
Fast One-Stage Unsupervised Domain Adaptive Person Search
Tianxiang Cui (Dalian Maritime University), Yang Wang (Hefei University of Technology)
RecognitionObject DetectionDomain AdaptationConvolutional Neural NetworkContrastive LearningImageVideo
🎯 What it does: This paper proposes a one-stage, end-to-end unsupervised domain adaptation person search framework named FOUS, which simultaneously performs detection and identity recognition in unlabeled target domains.
Fast Unpaired Multi-view Clustering
Xingfeng Li (Southwest University Of Science And Technology), Zhenwen Ren (Southeast University)
OptimizationComputational EfficiencyRepresentation LearningImageMultimodality
🎯 What it does: This paper proposes the Fast Unpaired Multi-view Clustering (FUMC) framework to achieve fast clustering on unpaired multi-view large-scale data.
Faster Optimal Coalition Structure Generation via Offline Coalition Selection and Graph-Based Search
Redha Taguelmimt (Univ Lyon), Tuomas Sandholm (Carnegie Mellon University)
OptimizationTabular
🎯 What it does: Proposed a new optimal coalition structure generation algorithm called SMART, which can efficiently solve cooperative allocation problems in multi-agent systems.
FastScene: Text-Driven Fast Indoor 3D Scene Generation via Panoramic Gaussian Splatting
Yikun Ma (Sun Yat-sen University), Zhi Jin (Sun Yat-sen University)
RestorationGenerationData SynthesisDepth EstimationVision Language ModelDiffusion modelGenerative Adversarial NetworkGaussian SplattingImageTextPoint Cloud
🎯 What it does: Propose the FastScene framework, enabling the rapid generation of high-quality, globally consistent 3D indoor scenes based solely on text prompts. The entire process includes panoramic generation, depth estimation, coarse view synthesis, progressive panoramic inpainting, and reconstruction based on 3D Gaussian splatting.
FBLG: A Local Graph Based Approach for Handling Dual Skewed Non-IID Data in Federated Learning
Yi Xu (Anhui University), Xiao Liu (Deakin University)
Federated LearningGraph Neural NetworkImage
🎯 What it does: Propose a federated learning algorithm called FBLG based on a local graph, specifically designed to handle non-IID data with dual skewness in label distribution and sample size.
Feature Norm Regularized Federated Learning: Utilizing Data Disparities for Model Performance Gains
Ke Hu (Shanghai Jiao Tong University), Weidong Qiu (Shanghai Jiao Tong University)
OptimizationFederated LearningConvolutional Neural NetworkImage
🎯 What it does: Proposed a federated learning algorithm called FNR-FL based on feature norm regularization, aiming to address model update bias caused by non-i.i.d. data.
FedConPE: Efficient Federated Conversational Bandits with Heterogeneous Clients
Zhuohua Li (Chinese University of Hong Kong), John C. S. Lui
Recommendation SystemOptimizationFederated LearningTextTabular
🎯 What it does: Propose FedConPE, a phased elimination-based federated conversational bandit algorithm that addresses collaborative learning and conversation strategies in heterogeneous multi-client environments
Federated Adaptation for Foundation Model-based Recommendations
Chunxu Zhang (Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education), Bo Yang (Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education)
Recommendation SystemFederated LearningSafty and PrivacyKnowledge DistillationTransformerTabular
🎯 What it does: Proposed a federated recommendation framework called FedPA based on foundational models, which uses pre-trained large-scale models as a shared knowledge base. It achieves user personalization by learning lightweight low-rank adapters only at the client side, and fuses general knowledge with personalized features through an adaptive gating mechanism.
Federated Multi-View Clustering via Tensor Factorization
Wei Feng (Xi'an Jiaotong University), Quanxue Gao (Xidian University)
Federated LearningMultimodality
🎯 What it does: Proposed a federated multi-view clustering framework named TensorFMVC, which utilizes tensor decomposition to achieve cross-client clustering learning.
Federated Prompt Learning for Weather Foundation Models on Devices
Shengchao Chen (University of Technology Sydney), Chengqi Zhang (University of Technology Sydney)
Federated LearningTransformerPrompt EngineeringTime Series
🎯 What it does: Proposed a federated learning framework called FedPoD for implementing personalized weather forecasting models on devices.
FedES: Federated Early-Stopping for Hindering Memorizing Heterogeneous Label Noise
Bixiao Zeng (Institute of Computing Technology, Chinese Academy of Sciences), Yingwei Zhang (Institute of Computing Technology, Chinese Academy of Sciences)
Federated LearningImage
🎯 What it does: Propose the FedES framework, which first estimates the noise rates of each client using signed Earth Mover's Distance, then achieves parameter-adaptive early stopping and noise-aware aggregation based on the estimation results, thereby enhancing robustness to heterogeneous noisy labels in federated learning.