AAAI 2026 Papers — Page 3
AAAI Conference on Artificial Intelligence · 4149 papers
Any2RSI: Controllable Remote Sensing Text-to-Image Generation via Any Control and Enriched Description
Xu Zhang (Wuhan University), Lefei Zhang (Wuhan University)
GenerationData SynthesisTransformerLarge Language ModelPrompt EngineeringVision Language ModelDiffusion modelContrastive LearningMultimodalityBenchmark
🎯 What it does: Propose the Any2RSI framework, enabling the flexible combination of multiple spatial controls for remote sensing text-to-image generation;
anyECG-chat: A Generalist ECG-MLLM for Flexible ECG Input and Multi-Task Understanding
Haitao Li (Zhejiang University), Zhengxing Huang (Zhejiang University)
SegmentationGenerationTransformerLarge Language ModelSupervised Fine-TuningContrastive LearningTime SeriesBiomedical DataElectrocardiogram
🎯 What it does: Constructed the anyECG dataset encompassing multiple tasks such as report generation, waveform localization, and multi-ECG comparison, and trained a multilingual large language model named anyECG-chat that supports dynamic length, lead reduction, and multi-ECG input.
AP2O-Coder: Adaptively Progressive Preference Optimization for Reducing Compilation and Runtime Errors in LLM-Generated Code
Jianqing Zhang (Shanghai Jiao Tong University), Jian Cao (Tencent)
OptimizationAI Code AssistantReinforcement Learning from Human FeedbackLarge Language ModelSupervised Fine-TuningReinforcement LearningText
🎯 What it does: Proposed a code generation enhancement method called AP2O-Coder based on offline preference optimization, significantly reducing compilation and runtime errors of LLM-generated code through generating exam-style code, automatic error analysis, step-by-step preference optimization, and adaptive error replay.
Aperiodic Tiling and Rhythmic Canons: A CP Journey
Guillaume Derval (University of Liège), Christophe Lecoutre (University of Artois and CNRS)
OptimizationComputational EfficiencyBenchmarkAudio
🎯 What it does: Modeling and solving the Aperiodic Tiling Complements Problem in music theory using constraint programming.
Apo2Mol: 3D Molecule Generation via Dynamic Pocket-Aware Diffusion Models
Xinzhe Zheng (University of Florida), Yanjun Li (University of Florida)
Drug DiscoveryProtein Structure PredictionGraph Neural NetworkDiffusion modelBiomedical Data
🎯 What it does: Propose a diffusion-based all-atom model called Apo2Mol, which can simultaneously generate ligands and corresponding holo pockets from apo protein structures.
Appearance Discrepancy-guided Sequence Hybrid Masking for Robust Scene Text Recognition
Shihao Zou (Huazhong University of Science and Technology), Wenfeng Xie (Ping An Property & Casualty Insurance Company of China, Ltd)
RecognitionTransformerImage
🎯 What it does: Proposes a sequence hybrid mask pre-training framework DSHM based on appearance difference guidance to enhance the robustness of scene text recognition.
Appearance-Motion Decomposed Alignment for Text-Video Retrieval
Meng Meng (Innovation Research Institute, Sangfor Technologies Inc.), Xu Zhou (Innovation Research Institute, Sangfor Technologies Inc.)
RetrievalTransformerVision Language ModelContrastive LearningImageVideoTextMultimodality
🎯 What it does: Propose an AMD-Net network for text-video retrieval, separating the spatial appearance and temporal motion features of videos to enhance the quality of cross-modal matching representations.
Approximately Envy-free and Equitable Allocations of Indivisible Items for Non-monotone Valuations
Vittorio Bilò (University of Salento), Cosimo Vinci (University of Salento)
OptimizationComputational Efficiency
🎯 What it does: Proposed and proved the existence of approximate fair allocations (EF1cg and EQ1cg) that satisfy the 'remove only one good and one bad' constraint under non-monotonic (non-negative or non-positive) values, and further extended the results under path constraints and target value scenarios.
Approximation Algorithm for Constrained k-Center Clustering: A Local Search Approach
Chaoqi Jia (Rmit University), Jason Xue (Deakin University)
OptimizationTabular
🎯 What it does: Studied the k-center clustering problem with instance-level cannot-link (CL) and must-link (ML) constraints, proposed a threshold and local search based dominating matching set (DMS) framework, and proved it can achieve optimal 2-approximation;
APT: Affine Prototype-Timestamp for Time Series Forecasting Under Distribution Shift
Yujie Li (Chinese Academy of Sciences), Fei Wang (Chinese Academy of Sciences)
Domain AdaptationRepresentation LearningTransformerTime SeriesBenchmark
🎯 What it does: Propose a lightweight Affine Prototype-Timestamp (APT) plugin to inject global distribution features into time series prediction, mitigating errors caused by distribution shifts.
APVR: Hour-Level Long Video Understanding with Adaptive Pivot Visual Information Retrieval
Hong Gao (Southeast University), Min-Ling Zhang (Southeast University)
RetrievalCompressionTransformerLarge Language ModelVision Language ModelVideoMultimodalityBenchmarkRetrieval-Augmented Generation
🎯 What it does: Propose the APVR framework to understand hour-level long videos without additional training, with the core idea of dual-layer retrieval: first, using Pivot Frame Retrieval (PFR) to evaluate frame-level confidence and perform adaptive iterative sampling through query expansion, CLIP, and Grounding-DINO; then, using Pivot Token Retrieval (PTR) to compress and select the most important visual tokens at the token level of the visual encoder through query-aware attention scoring, dynamic chunking, and head-level soft voting, ultimately providing multi-modal large language models (MLLM) for reasoning.
AquaSplatting: A Hybrid 3D Representation for Robust Underwater Scene Reconstruction via Dual-Branch Rendering
Jiangbei Hu (Dalian University of Technology), Ying He (Dalian University of Technology)
RestorationDepth EstimationComputational EfficiencyGaussian SplattingImage
🎯 What it does: Propose AquaSplatting dual-branch hybrid 3D representation for real-time reconstruction of underwater scenes and recovery of non-water images.
AR-Nav Benchmark: Augmented Reality Navigation with Vision and Language
Liqi Yan (Hangzhou Dianzi University), Pan Li (Hangzhou Dianzi University)
OptimizationComputational EfficiencyData-Centric LearningGraph Neural NetworkTransformerLarge Language ModelVision Language ModelSimultaneous Localization and MappingImageVideoTextMultimodalityBenchmark
🎯 What it does: Constructed the AR-Nav benchmark, proposed AR-VLM² for generating structured visual-language memory, designed ARN-Pilot for rapid target retrieval and path planning, and implemented a dual-modal interactive synthesis module, enabling AR navigation to achieve long-term spatiotemporal memory, semantic understanding, and real-time personalized guidance.
ARBench: Algorithmic Reasoner or API Alchemist? Evaluating LLMs Beyond API Calls
Ren-Biao Liu (Nanjing University), Ming Li (Nanjing University)
Large Language ModelTextBenchmark
🎯 What it does: This paper proposes the ARBench benchmark to evaluate the algorithmic reasoning capabilities of large language models in implementing machine learning algorithms from scratch without using high-level APIs.
Arbitrary-Scale 3D Gaussian Super-Resolution
Huimin Zeng (Northeastern University), Yun Fu (Northeastern University)
Super ResolutionDiffusion modelGaussian SplattingImageBenchmark
🎯 What it does: Propose an arbitrary-scale high-resolution view rendering framework based on a single 3D Gaussian model, capable of generating high-quality images at both integer and non-integer upscaling ratios.
ARCHE: A Novel Task to Evaluate LLMs on Latent Reasoning Chain Extraction
Pengze Li (Artificial Intelligence Innovation and Incubation Institute of Fudan University), Xi Chen (Artificial Intelligence Innovation and Incubation Institute of Fudan University)
TransformerLarge Language ModelPrompt EngineeringMixture of ExpertsTextGraphBenchmarkRetrieval-Augmented Generation
🎯 What it does: Proposes the Latent Reasoning Chain Extraction (ARCHE) task, requiring large language models to extract and structure implicit reasoning chains from scientific paper introductions, presenting them in the form of Reasoning Logic Tree (RLT), with nodes as viewpoints and edges as refined forms of Peirce's three types of reasoning (deductive, inductive, abductive);
ArchetypeTrader: Reinforcement Learning for Selecting and Refining Learnable Strategic Archetypes in Quantitative Trading
Chuqiao Zong (Nanyang Technological University), Bo An (Nanyang Technological University)
Reinforcement LearningAuto EncoderTime SeriesFinance Related
🎯 What it does: Propose the ArchetypeTrader framework, which generates single-transaction demonstrations using dynamic programming, refines discrete trading archetypes through a VQ encoder-decoder, and then uses RL to select appropriate archetypes, achieving refinement of archetypes through one-time strategy adjustments within each time period.
ArchRAG: Attributed Community-based Hierarchical Retrieval-Augmented Generation
Shu Wang (Chinese University of Hong Kong), Yuchi Ma (Huawei Cloud Computing Technologies Company Limited)
GenerationRetrievalComputational EfficiencyLarge Language ModelTextGraphRetrieval-Augmented Generation
🎯 What it does: Proposed a hierarchical retrieval-enhanced generation method called ArchRAG based on attribute communities, which efficiently retrieves and injects graph-structured knowledge into large language models to complete question-answering tasks.
ARDiff: Anisotropic Residual Diffusion for Heterogeneous Graph Learning
Yong Chen (Xiamen University), Song-Zhi Su (Xiamen University)
ClassificationRecommendation SystemGraph Neural NetworkDiffusion modelGraph
🎯 What it does: Proposed and implemented ARDiff, a residual diffusion framework for heterogeneous graph learning, which can progressively denoise and extract relational semantics on the semantic chain CoS.
Are Graph Transformers Necessary? Efficient Long-Range Message Passing with Fractal Nodes in MPNNs
Jeongwhan Choi (Korea Advanced Institute of Science & Technology), Noseong Park (Korea Advanced Institute of Science & Technology)
Computational EfficiencyRepresentation LearningGraph Neural NetworkGraph
🎯 What it does: Proposes a fractal nodes mechanism, which divides the graph into subgraphs and introduces nodes representing subgraphs on the original graph, directly aggregating information at the subgraph level and mutually updating with original nodes.
Are Language Models Any Good at Density Modeling?
Sriram Ranga (Nanyang Technological University), Anupam Chattopadhyay (Nanyang Technological University)
Data SynthesisTransformerLarge Language ModelText
🎯 What it does: This paper constructs a synthetic language where numbers are written as English words, trains a GPT-2 structure autoregressive model, and evaluates the model's density modeling capability under various probability distributions (uniform, linear, binomial, sine);
Are We on the Right Way to Assess Document Retrieval-Augmented Generation?
Wenxuan Shen (South China University of Technology), Weiwei Lin (Huazhong University of Science and Technology)
RetrievalTransformerLarge Language ModelTextMultimodalityBenchmarkRetrieval-Augmented Generation
🎯 What it does: Proposed a large-scale, multimodal, multilingual document retrieval-augmented generation evaluation benchmark called DOUBLE-BENCH, conducting fine-grained evaluation of retrieval and generation components on it.
Area-Optimal Control Strategies for Heterogeneous Multi-Agent Pursuit
Kamal Mammadov (University of Adelaide), Damith C. Ranasinghe (University of Adelaide)
Optimization
🎯 What it does: Proposes a multi-agent pursuit-evasion control strategy based on minimizing the area of the safe reachable set, defining the safe reachable set as the intersection of Apollonius circles between each pursuer-escaper pair and providing an immediate optimal direction.
Argumentative Debates for Transparent Bias Detection
Hamed Ayoobi (University of Groningen), Francesca Toni (Imperial College London)
Explainability and InterpretabilityLarge Language ModelTextTabular
🎯 What it does: Propose a transparent bias detection framework ABIDE based on argumentative debate, which judges individual and group bias situations by constructing a quantized bipolar argumentation framework (QBAF) that leverages neighborhood information.
ARIW-Framework: Adaptive Robust Iterative Watermarking Framework
Shaowu Wu (Sun Yat-sen University), Wei Lu (Sun Yat-sen University)
Safty and PrivacyImage
🎯 What it does: Proposed an Adaptive Robust Iterative Watermarking Framework (ARIW-Framework), achieving dual goals of high visual quality and robustness through residual iterative optimization, parallel noise layers, and gradient-adaptive embedding.
ARNS: Adaptive Relation-Aware Negative Sampling with Curriculum Learning for Inductive Knowledge Graph Completion
Ling Ding (Tianjin University), Lei Huang (Tianjin University)
Representation LearningGraph Neural NetworkContrastive LearningGraph
🎯 What it does: Propose ARNS, an adaptive relation-aware negative sampling method based on GNN, for improving inductive knowledge graph completion.
ARQUSUMM: Argument-aware Quantitative Summarization of Online Conversations
An Quang Tang (RMIT University), Zhuang Li (RMIT University)
GenerationTransformerLarge Language ModelText
🎯 What it does: An adaptive argument-aware quantitative summarization framework, ARQUSUMM, is proposed for dialogues on online discussion platforms, achieving structured extraction and quantification of argument claims and reasons.
ARTEM: Enhancing Large Language Model Agents with Spatial-Temporal Episodic Memory
Cassandra Hui-Ming Tan (Singapore Management University), Ah-Hwee Tan (Singapore Management University)
RetrievalRepresentation LearningTransformerLarge Language ModelTextBenchmark
🎯 What it does: Propose the ARTEM architecture, integrating large language models with a self-organizing spatiotemporal episodic memory network (STEM) to achieve more precise event encoding and retrieval.
ASCD: Attention-Steerable Contrastive Decoding for Reducing Hallucination in MLLM
Yujun Wang, Yunpu Ma (Christian Albrechts University Of Kiel)
TransformerVision Language ModelContrastive LearningMultimodalityBenchmark
🎯 What it does: This paper proposes an attention scheduling-based contrastive decoding method (ASCD), which significantly reduces visual hallucinations and enhances visual-semantic consistency by directly adjusting the attention distribution of multi-modal large language models during inference.
AsFT: Anchoring Safety During LLM Fine-Tuning Within Narrow Safety Basin
Shuo Yang (Peking University), Li Yuan (Peking University)
OptimizationSafty and PrivacyTransformerLarge Language ModelSupervised Fine-TuningText
🎯 What it does: Propose a safety-guided fine-tuning method called AsFT, which maintains model safety during fine-tuning by projecting parameter updates into a subspace aligned with the alignment direction (safety direction) and suppressing updates orthogonal to it;
ASKD: Reinforcement Learning-Style Knowledge Distillation with Quality-Adaptive Skewness
Mingjie Zhang (Peking University), Wei Ye (Peking University)
Knowledge DistillationReinforcement LearningText
🎯 What it does: Proposed ASKD, a reinforcement learning-based knowledge distillation framework that enhances the learning effectiveness of student models through quality-adaptive skewness.
Assemble Your Crew: Automatic Multi-agent Communication Topology Design via Autoregressive Graph Generation
Shiyuan Li (Griffith University), Shirui Pan (Griffith University)
GenerationOptimizationRecurrent Neural NetworkGraph Neural NetworkLarge Language ModelSupervised Fine-TuningAgentic AITextGraphBenchmark
🎯 What it does: Propose ARG-DESIGNER, a self-recursive graph generation model from scratch, to automatically design communication topologies for multi-agent systems (MAS), generating the required number of agents, roles, and communication links based on natural language task queries.
Assessing LLMs for Serendipity Discovery in Knowledge Graphs: A Case for Drug Repurposing
Mengying Wang (Case Western Reserve University), Yinghui Wu (Case Western Reserve University)
Drug DiscoveryGraph Neural NetworkTransformerLarge Language ModelPrompt EngineeringGraphBiomedical DataBenchmark
🎯 What it does: Designed the SerenQA evaluation framework, constructed a drug repurposing QA benchmark based on the Clinical Knowledge Graph, and systematically evaluated the performance of LLMs in discovering 'serendipity' answers.
Assessing the Capabilities of LLMs in Humor: A Multi-dimensional Analysis of Oogiri Generation and Evaluation
Ritsu Sakabe (Hitotsubashi University), Mamoru Komachi (Hitotsubashi University)
GenerationData SynthesisTransformerLarge Language ModelImageText
🎯 What it does: The study constructs a multi-dimensional humor generation and evaluation dataset based on the Japanese improvisational comedy game 'Oogiri,' systematically assessing the performance of LLMs in generating humorous content and evaluating humor, while exploring differences in humor perception between humans and LLMs.
Assignment Problems in Cost Function Networks
Guidio Sewa (MIAT, INRAE), Thomas Schiex (MIAT, INRAE)
OptimizationBenchmark
🎯 What it does: This paper proposes and implements a soft local consistency propagator based on the linear assignment problem for efficiently handling ALLDIFFERENT constraints in weighted constraint satisfaction problems.
ASSIST-3D: Adapted Scene Synthesis for Class-Agnostic 3D Instance Segmentation
Shengchao Zhou (University of Hong Kong), Xiaojuan Qi (University of Hong Kong)
SegmentationGenerationData SynthesisLarge Language ModelSupervised Fine-TuningPoint Cloud
🎯 What it does: Propose ASSIST-3D, a synthetic scene generation pipeline specifically designed for class-agnostic 3D instance segmentation, capable of generating realistic point cloud data under three principles: geometric diversity, contextual complexity, and layout plausibility, thereby enhancing model generalization capabilities.
AStar: Boosting Multimodal Reasoning with Automated Structured Thinking
Jinyang Wu (Tsinghua University), Jianhua Tao (Tsinghua University)
Explainability and InterpretabilityComputational EfficiencyLarge Language ModelVision Language ModelImageTextMultimodalityRetrieval-Augmented GenerationChain-of-Thought
🎯 What it does: Proposed a training-free, multimodal reasoning framework called AStar based on automated structured thinking. It provides external explicit reasoning guidance to models through 'thought cards,' helping LLMs combine internal implicit capabilities during reasoning to achieve more efficient and accurate multimodal reasoning.
Asymmetric Cross-Modal Knowledge Distillation: Bridging Modalities with Weak Semantic Consistency
Riling Wei (Zhejiang Laboratory), Chao Li (Chinese Academy of Sciences)
Knowledge DistillationContrastive LearningMultimodalityBenchmark
🎯 What it does: In remote sensing scenarios where multispectral and RGB images lack strong semantic correspondence, the heterogeneous cross-modal knowledge distillation (ACKD) method is proposed, along with the SemBridge framework to enable knowledge transfer.
Asymptotic and Finite Sample Analysis of Nonexpansive Stochastic Approximations with Markovian Noise
Ethan Blaser (University of Virginia), Shangtong Zhang (University of Virginia)
OptimizationTabular
🎯 What it does: Studied the asymptotic and finite-sample analysis of non-expansive stochastic approximation with Markov noise, particularly focusing on stochastic approximation algorithms under non-contractive operators.
AT-Field: Rethinking the Games in Adversarial Training
Yixiao Xu (Beijing University of Posts and Telecommunications), Zhihong Tian (Guangzhou University)
ClassificationAdversarial AttackImage
🎯 What it does: Propose the AT-Field framework, which dynamically reorganizes training batches by leveraging game-theoretic relationships between samples, transforming non-potential games into potential games;
Attack the Messages, Not the Agents: A Multi-round Adaptive Stealthy Tampering Framework for LLM-MAS
Bingyu Yan (Beihang University), Litian Zhang (Beihang University)
Adversarial AttackTransformerLarge Language ModelReinforcement LearningTextBenchmark
🎯 What it does: Propose a multi-round adaptive stealthy tampering framework, MAST, designed to attack the communication环节 of large language model multi-agent systems (LLM-MAS), capable of altering system outputs without being detected.
Attention Retention for Continual Learning with Vision Transformers
Yue Lu (Northwestern Polytechnical University), Wencong Zhang (Northwestern Polytechnical University)
TransformerSupervised Fine-TuningImageBenchmark
🎯 What it does: Propose an attention retention framework on Vision Transformers, utilizing gradient masking to suppress attention drift, thereby mitigating catastrophic forgetting in continual learning.
Attention to Threat-Relevant Objects: Reasoning Detection in Autonomous Driving via Multimodal Large Language Models
Yulin He (National University of Defense Technology), Yusong Tan (National University of Defense Technology)
Autonomous DrivingTransformerReinforcement LearningVision Language ModelImageVideoMultimodalityBenchmarkChain-of-Thought
🎯 What it does: Proposed the threat-oriented reasoning detection task and built the corresponding benchmark; designed two methods, ThreatCoT (untrained CoT + visual toolchain) and ThreatReasoner (end-to-end reasoning based on GRPO), to enhance threat object detection and threat level assessment.
Attentive Keypoint Identification: Progressive Spatiotemporal Refinement for Video-based Human Pose Estimation
Sifan Wu (Jilin University), Yingying Jiao (Zhejiang University of Technology)
Pose EstimationTransformerVideo
🎯 What it does: Proposes PSTPose, a framework that enhances video human pose estimation through phased spatial-temporal refinement, primarily addressing issues of spatial uniformity and temporal rigidity.
Attribute-guided Dynamic Prompt Learning for Graph Neural Networks
Zhuomin Liang, Xian Yang (University of Manchester)
Representation LearningAdversarial AttackGraph Neural NetworkPrompt EngineeringContrastive LearningGraph
🎯 What it does: Generate dynamic prompt vectors using node attributes to enhance model robustness under structural perturbations while keeping the original GNN parameters unchanged.
Attribution Analysis-based Concept Alignment: A Human-in-the-loop Data Debugging Framework
Lei Chai (Beihang University), Jingxuan Xu (Beihang University)
Explainability and InterpretabilityData-Centric LearningDiffusion modelImageText
🎯 What it does: Proposed the AACA framework, which performs human-computer interactive image data debugging through attribute analysis and concept alignment;
Audio-Assisted Face Video Restoration with Temporal and Identity Complementary Learning
Yuqin Cao (Shanghai Jiao Tong University), Xiongkuo Min (Shanghai Jiao Tong University)
RestorationSuper ResolutionConvolutional Neural NetworkVideoMultimodalityAudio
🎯 What it does: Propose a generic audio-assisted facial video restoration network called GAVN, which can simultaneously handle three types of distortions: video compression artifacts, blurriness, and super-resolution.
Audio-Thinker: Guiding Large Audio Language Model When and How to Think via Reinforcement Learning
Shu Wu, Dong Yu (Tencent AI Lab)
Large Language ModelReinforcement LearningMultimodalityChain-of-ThoughtAudio
🎯 What it does: Proposed a reinforcement learning framework called Audio-Thinker to guide large audio language models (LALM) on when and how to perform deep reasoning.
Augmentation-invariant Learning Strategy via Data Augmentation for Improving Model Generalization
Yu Miao, Yan Qiang (Taiyuan University Of Technology)
ClassificationBiomedical Data
🎯 What it does: Proposed the AILS implicit incremental learning strategy, generating semantically invariant data augmentation in the feature space through variational Bayesian inference to enhance the generalization ability of medical image classification.
Augmented Runtime Collaboration for Self-Organizing Multi-Agent Systems: A Hybrid Bi-Criteria Routing Approach
Qingwen Yang (Northeastern University), Yingyou Wen (Northeastern University)
OptimizationComputational EfficiencyData-Centric LearningTransformerLarge Language ModelAgentic AITextBenchmark
🎯 What it does: Propose BiRouter, a dual-standard (importance + coherence) routing method that enables each agent in self-organizing multi-agent systems to adaptively select the next hop based solely on local information, forming globally efficient task chains.
Augmenting Intra-Modal Understanding in MLLMs for Robust Multimodal Keyphrase Generation
Jiajun Cao (Xiamen University), Jinsong Su (Xiamen University)
GenerationTransformerLarge Language ModelSupervised Fine-TuningVision Language ModelImageTextMultimodality
🎯 What it does: This paper proposes the AimKP framework to improve multimodal key phrase generation tasks, significantly enhancing the model's cross-modal and single-modal understanding capabilities.
AURORA: Augmented Understanding via Structured Reasoning and Reinforcement Learning for Reference Audio-Visual Segmentation
Ziyang Luo, Junwei Han (Mohamed bin Zayed University of Artificial Intelligence)
SegmentationReinforcement LearningMultimodalityChain-of-Thought
🎯 What it does: Proposes a new framework called AURORA, aiming to enhance language understanding and reasoning capabilities in reference audio-visual segmentation (Ref-AVS) tasks through structured reasoning and reinforcement learning.
Authority Backdoor: A Certifiable Backdoor Mechanism for Authoring DNNs
Han Yang (Southeast University), Zhen Ling (Southeast University)
Safty and PrivacyConvolutional Neural NetworkTransformerImage
🎯 What it does: Proactively lock the DNN after model leakage, enabling it to function normally only under specific hardware triggers.
AuthSig: Safeguarding Scanned Signatures Against Unauthorized Reuse in Paperless Workflows
Ruiqiang Zhang (University of Science and Technology of China), Weiming Zhang (University of Science and Technology of China)
Safty and PrivacyDiffusion modelAuto EncoderImage
🎯 What it does: Propose AuthSig, a static electronic signature framework based on generative models and watermarks, achieving one-time use and document-bound signature verification.
Auto-PRE: An Automatic and Cost-Efficient Peer-Review Framework for Language Generation Evaluation
Junjie Chen (Tsinghua University), Qingyao Ai (Ant Group)
GenerationComputational EfficiencyTransformerLarge Language ModelText
🎯 What it does: Propose Auto-PRE, an LLM evaluation framework based on automated qualification exams, which screens LLMs using three key metrics: consistency, relevance, and confidence;
AutoGameUI: Constructing High-Fidelity GameUI via Multimodal Correspondence Matching
Zhongliang Tang (Tencent), Fei Xia (Tencent)
GenerationGraph Neural NetworkTransformerContrastive LearningImageMultimodality
🎯 What it does: Built an automated system called AutoGameUI, leveraging multimodal learning and hierarchical correspondence matching to high-fidelity combine UI and UX design into complete game UIs; simultaneously provides an interactive Web tool to support human-AI collaborative editing.
AutoLink: Autonomous Schema Exploration and Expansion for Scalable Schema Linking in Text-to-SQL at Scale
Ziyang Wang (Huazhong University of Science and Technology), Xiang Bai (Huazhong University of Science and Technology)
AI Code AssistantTransformerLarge Language ModelAgentic AITextTabularRetrieval-Augmented Generation
🎯 What it does: Propose AutoLink, an autonomous agent framework based on large language models (LLMs), transforming schema linking into an iterative, interactive exploration process;
AutoMalDesc: Large-Scale Script Analysis for Cyber Threat Research
Alexandru-Mihai Apostu (CrowdStrike), Mihaela Gaman (University of Bucharest)
Anomaly DetectionTransformerLarge Language ModelSupervised Fine-TuningText
🎯 What it does: Propose AutoMalDesc, an automated framework that generates large-scale script security descriptions and detects malicious scripts by leveraging a small number of expert-labeled samples through iterative self-supervised learning.
Automata-less Monitoring via Trace-Checking
Andrea Brunello (University of Udine), Nicola Saccomanno (University of Udine)
Computational Efficiency
🎯 What it does: This paper proves that within specific LTL/LTLf fragments, runtime monitoring can be accomplished without constructing a DFA, directly using trace-checking.
Automated Human Strategic Behavior Modeling via Large Language Models
Xiaohan Xie (Chinese University of Hong Kong), Jianwei Huang (Chinese University of Hong Kong)
OptimizationExplainability and InterpretabilityTransformerLarge Language Model
🎯 What it does: Proposes the AutoBM framework, which leverages large language models to automatically generate, evaluate, and refine interpretable behavioral models;
Automated Repair of Totally-Ordered Hierarchical Task Network Domains via Context-Free Grammars with Large Language Model Support
Daniel Lutalo (Australian National University), Pascal Bercher (Australian National University)
OptimizationComputational EfficiencyAI Code AssistantTransformerLarge Language ModelPrompt EngineeringGraph
🎯 What it does: Propose a technique that leverages large language models (LLM) to repair missing actions in the domain model of totally ordered hierarchical task networks (TO-HTN), converting the repair problem into a context-free grammar (CFG) repair task, and using structural filtering and masked prompts to ensure the LLM operates solely at the symbolic level;
Automatic Channel Pruning by Searching with Structure Embedding for Hash Network
Zifan Liu, Heng Qi (Dalian University of Technology)
RetrievalCompressionGraph Neural NetworkContrastive LearningImage
🎯 What it does: Propose ACP-SSE, an automatic channel pruning framework based on graph structure embedding for deep hash networks.
Automatic Paper Reviewing with Heterogeneous Graph Reasoning over LLM-Simulated Reviewer-Author Debates
Shuaimin Li (Chinese Academy of Sciences), Min Yang (Chinese Academy of Sciences)
ClassificationGraph Neural NetworkTransformerLarge Language ModelPrompt EngineeringTextGraph
🎯 What it does: Propose the ReViewGraph framework, which uses multi-agent LLMs to simulate multi-round debates between reviewers and authors, extracts viewpoints and debate relationships, constructs a heterogeneous debate graph, and predicts paper acceptance/rejection decisions using graph neural network reasoning.
Automatic Translational Correction of Multi-View Coronary Angiography Based on Auto-Annotation Data Generation
Yue Cao (Tianjin University), Jiachen Yang (Tianjin University)
GenerationData SynthesisOptimizationConvolutional Neural NetworkGenerative Adversarial NetworkImageBiomedical Data
🎯 What it does: Propose a fully automated multi-view coronary angiography translational correction framework without manual annotation, achieving high-precision correction using generative models and evolutionary algorithms.
Automating Complex Document Workflows via Stepwise and Rollback-Enabled Operation Orchestration
Yanbin Zhang (East China Normal University), Renjun Hu (East China Normal University)
ClassificationTransformerLarge Language ModelTextMultimodalitySequentialBenchmark
🎯 What it does: Proposed the AutoDW framework, achieving multi-step workflow automation for document processing, supporting incremental planning and rollback;
Autonomous Concept Drift Threshold Determination
Pengqian Lu (University of Technology Sydney), Guangquan Zhang (University of Technology Sydney)
Anomaly DetectionComputational EfficiencyImageTabularTime SeriesBenchmark
🎯 What it does: Proposed and implemented the Dynamic Threshold Determination (DTD) algorithm to automatically adjust the threshold of concept drift detectors, thereby improving the overall predictive performance of the model.
Autonomous Partner Selection for Cooperative Multi-Agent Reinforcement Learning
Rui Tang (Central South University), Yongzheng Cui (Central South University)
Recurrent Neural NetworkTransformerReinforcement LearningContrastive LearningBenchmark
🎯 What it does: Proposes the APS framework, enabling multiple agents to autonomously select partners and implicitly form groups under CTDE environments to enhance team collaboration performance.
Autonomous Vehicle Path Planning by Searching with Differentiable Simulation
Asen Nachkov (INSAIT Sofia University St Kliment Ohridski), Luc Van Gool (INSAIT Sofia University St Kliment Ohridski)
Autonomous DrivingOptimizationWorld Model
🎯 What it does: Proposed a framework for path planning using differentiable simulation during testing, named Differentiable Simulation for Search (DSS), which optimizes the action sequence of autonomous vehicles by performing gradient search on future trajectories within the differentiable simulator Waymax;
AutoPP: Towards Automated Product Poster Generation and Optimization
Jiahao Fan (JD.COM), Ching Law (JD.COM)
GenerationOptimizationTransformerLarge Language ModelReinforcement LearningPrompt EngineeringVision Language ModelDiffusion modelImageText
🎯 What it does: Built the AutoPP automated pipeline to achieve end-to-end generation of high-quality product posters from basic product information, with iterative optimization through online CTR feedback.
AutoSchA: Automatic Hierarchical Music Representations via Multi-Relational Node Isolation
Stephen Ni-Hahn (Duke University), Simon Mak (Stanford University)
Representation LearningGraph Neural NetworkTransformerGraph
🎯 What it does: This paper proposes AutoSchA, an automatic hierarchical music representation framework based on graph neural networks, for generating Scheneckria analyses comparable to those of human experts.
AutoTool: Efficient Tool Selection for Large Language Model Agents
Jingyi Jia (Huazhong University of Science and Technology), Qinbin Li (Huazhong University of Science and Technology)
Computational EfficiencyGraph Neural NetworkTransformerLarge Language ModelAgentic AIContrastive LearningTextGraph
🎯 What it does: This paper proposes the AutoTool framework, which constructs a graph structure using tool usage inertia to enable LLM-call-free selection of tools and parameters within an LLM agent.
AUVIC: Adversarial Unlearning of Visual Concepts for Multi-modal Large Language Models
Haokun Chen (LMU Munich), Volker Tresp (University of Oxford)
Adversarial AttackTransformerLarge Language ModelGenerative Adversarial NetworkContrastive LearningMultimodalityBenchmark
🎯 What it does: Propose the AUVIC framework, achieving precise forgetting of visual concepts in multimodal large language models through adversarial perturbations on visual inputs;
Auxiliary Gene Learning: Spatial Gene Expression Estimation by Auxiliary Gene Selection
Kaito Shiku (Kyushu University), Ryoma Bise (Kyushu University)
OptimizationImageBiomedical Data
🎯 What it does: By treating neglected low-expressed genes as auxiliary tasks, joint training is employed to enhance the accuracy of predicting spatial gene expression from pathological images.
AV-Edit: Multimodal Generative Sound Effect Editing via Audio-Visual Semantic Joint Control
Xinyue Guo (Xiaomi Inc.), Jian Luan (Xiaomi Inc.)
GenerationData SynthesisTransformerDiffusion modelAuto EncoderContrastive LearningVideoTextMultimodalityAudio
🎯 What it does: Propose AV-Edit, a fine-grained, semantically aligned sound effect editing framework that leverages visual, audio, and text multimodal information to add, delete, or replace audio elements in videos.
AV-SSAN: Audio-Visual Selective DOA Estimation Through Explicit Multi-Band Semantic-Spatial Alignment
Yu Chen (University of Science and Technology Beijing), Xinyuan Qian (Fano)
RecognitionConvolutional Neural NetworkTransformerVision Language ModelMultimodalityAudio
🎯 What it does: Propose a cross-instance visual cue-based audio-visual sound source localization task (CI-AVL) and a corresponding multi-band semantic space alignment framework (AV-SSAN) to achieve selective localization of target sound sources.
AVM: Towards Structure-Preserving Neural Response Modeling in the Visual Cortex Across Stimuli and Individuals
Qi Xu (Dalian University of Technology), Yangfan Hu (Zhejiang University of Finance and Economics)
TransformerBiomedical Data
🎯 What it does: Propose the Adaptive Visual Model (AVM), achieving structural preservation and functional adaptation under visual stimuli, individual differences, and environmental changes by freezing the Vision Transformer encoder and inserting a lightweight conditional modulation module in each Transformer block.
Aware Distillation for Robust Vision-Language Tracking Under Linguistic Sparsity
Guangtong Zhang (Jilin University), Tian Bai (Guangxi Normal University)
Object TrackingComputational EfficiencyKnowledge DistillationRepresentation LearningTransformerVision Language ModelContrastive LearningImageTextMultimodality
🎯 What it does: Propose ADTrack, a lightweight vision-language tracking model achieved through knowledge distillation and contrastive learning, which maintains robustness and enhances localization performance in language-scarce scenarios.
Aware First, Think Less: Dynamic Boundary Self-Awareness Drives Significant Gains in Reasoning Efficiency in Large Language Models
Qiguang Chen (Harbin Institute of Technology), Wanxiang Che (Harbin Institute of Technology)
Computational EfficiencyLarge Language ModelReinforcement LearningTextBenchmark
🎯 What it does: Propose a Dynamic Reasoning Boundary Self-Aware Framework (DR.SAF), enabling large language models to assess problem difficulty in real-time during reasoning based on their current reasoning capabilities, dynamically adjusting reasoning depth and output length, thereby significantly reducing redundant reasoning steps.
Axis-Aligned Document Dewarping
Chaoyun Wang (Xi'an Jiaotong University), Caigui Jiang (University of Tokyo)
RestorationConvolutional Neural NetworkImage
🎯 What it does: This paper proposes a lightweight convolutional network based on axis-aligned geometric constraints for de-warping single document images, and introduces an axis-aligned preprocessing strategy along with a new evaluation metric called AAD.
Backdooring Rationalization
Lingxiao Kong (Huazhong University of Science and Technology), Lei Wu (Huazhong University of Science and Technology)
Explainability and InterpretabilityAdversarial AttackRecurrent Neural NetworkTransformerSupervised Fine-TuningPrompt EngineeringText
🎯 What it does: This paper systematically investigates, for the first time, the vulnerability of rationalization models to backdoor attacks, and proposes a specialized backdoor attack method called BadRNL, which can achieve high-success-rate backdoor control without significantly affecting normal performance.
Backtrace Mamba: Reviving Critical Temporal Contexts via Hierarchical Memory Compression for Online Action Detection
Su Yan (Xidian University), Cheng Deng (Xidian University)
Object DetectionComputational EfficiencyVideo
🎯 What it does: Proposed a Mamba-based online action detection framework called MOAD, which can effectively capture long-term temporal dependencies under real-time conditions.
BadThink: Triggered Overthinking Attacks on Chain-of-Thought Reasoning in Large Language Models
Shuaitong Liu (Southwest University), Gaojie Jin (University of Exeter)
Computational EfficiencyAdversarial AttackTransformerLarge Language ModelSupervised Fine-TuningPrompt EngineeringTextChain-of-Thought
🎯 What it does: Proposed and implemented a backdoor attack called BadThink during training, which can induce 'overthinking' in chain-of-thought (CoT) large language models, significantly increasing inference time and computational cost without altering the final answer.
BAG: Benchmarking Anomaly Detection on Dynamic Graphs
Fengrui Hua (IDEA Research, International Digital Economy Academy), Jian Guo (IDEA Research, International Digital Economy Academy)
Anomaly DetectionGraph Neural NetworkGraphBenchmark
🎯 What it does: Propose the BAG benchmark, systematically evaluating the performance of 25 DyGAD models on 10 real-world dynamic graph datasets.
Balanced Knowledge Distillation for Large Language Models with Mix-of-Experts
Jiajun Liu (Southeast University), Zijie Xu (Southeast University)
Knowledge DistillationLarge Language ModelMixture of ExpertsText
🎯 What it does: Proposes a new knowledge distillation framework called B-Distill under the MoE structure to compress large language models while maintaining the MoE architecture.
Balancing Multimodal Domain Generalization via Gradient Modulation and Projection
Hongzhao Li (Zhengzhou University), Muhammad Haris Khan (Mohamed Bin Zayed University of Artificial Intelligence)
Domain AdaptationGenerative Adversarial NetworkMultimodality
🎯 What it does: Proposed the Gradient Modulation Projection (GMP) strategy to achieve gradient balance and conflict resolution in multi-modal domain generalization training.
Balancing Signal and Variance: Adaptive Offline RL Post-Training for VLA Flow Models
Hongyin Zhang (Zhejiang University), Donglin Wang (Westlake University)
Robotic IntelligenceReinforcement LearningVision-Language-Action ModelFlow-based ModelMultimodality
🎯 What it does: Propose an adaptive offline reinforcement learning fine-tuning method, ARFM, which utilizes energy-weighted flow matching to fine-tune Vision-Language-Action (VLA) flow models, thereby improving the action accuracy in complex downstream tasks.
BAMAS: Structuring Budget-Aware Multi-Agent Systems
Liming Yang (Peking University), Zhenpeng Chen (Nanyang Technological University)
OptimizationLarge Language ModelReinforcement LearningAgentic AITextBenchmark
🎯 What it does: Designed the BAMAS framework, first selecting the optimal LLM set within the budget using integer linear programming, then choosing the collaborative topology via offline reinforcement learning, and finally instantiating a budget-aware multi-agent system.
Bandit Learning in Housing Markets
Shiyun Lin (Peking University)
Reinforcement LearningTabularFinance Related
🎯 What it does: Proposed a multi-player multi-armed bandit framework for unknown preference learning in housing markets, and introduced core regret as an evaluation metric;
Bant: Byzantine Antidote via Trial Function and Trust Scores
Gleb Molodtsov (Moscow Independent Research Institute of Artificial Intelligence), Aleksandr Beznosikov (Trusted Ai Research Center, Russian Academy Of Sciences)
Federated LearningImageTextBiomedical DataElectrocardiogram
🎯 What it does: Proposed Byzantine-robust federated learning algorithms Bant and AutoBant based on trial functions and trust scores, which can achieve convergence even when Byzantine nodes are in the majority.
BAT: Learning Event-based Optical Flow with Bidirectional Adaptive Temporal Correlation
Gangwei Xu (Huazhong University of Science and Technology), Xin Yang (Zhejiang University)
Convolutional Neural NetworkRecurrent Neural NetworkOptical FlowTime Series
🎯 What it does: This paper proposes the BAT framework, which estimates optical flow for event cameras by leveraging bidirectional adaptive temporal correlation.
Bayes-Optimal Fair Classification with Multiple Sensitive Features
Yi Yang, Xiangyu Chang (Xi'an Jiaotong University)
ClassificationTabular
🎯 What it does: This paper theoretically derives the Bayes-optimal fair classifier under scenarios with multiple sensitive features, and provides Bayes-optimal decision rules that satisfy approximate fairness metrics (mean difference MD and mean ratio MR). Subsequently, it proposes an in-processing method based on regularization during training and a post-processing method based on calibrated plug-in approaches, which can learn these Bayes-optimal classifiers from finite samples.
BayesAgent: Bayesian Agentic Reasoning Under Uncertainty via Verbalized Probabilistic Graphical Modeling
Hengguan Huang (University of Copenhagen), Samir Bhatt (University of Copenhagen)
Explainability and InterpretabilityTransformerLarge Language ModelAgentic AIPrompt EngineeringTextMultimodalityChain-of-Thought
🎯 What it does: Propose a method that uses natural language prompts to enable large language models (LLMs) to autonomously construct latent variables and their dependencies, leveraging numerical Bayesian inference and differentiable calibration loss to form the Verbalized Probabilistic Graphical Modeling (vPGM) framework, achieving agency reasoning and uncertainty quantification without requiring expert-designed models.
Bayesian Network Structural Consensus via Greedy Min-Cut Analysis
Pablo Torrijos (Universidad de Castilla-La Mancha), José A. Gámez (Universidad de Castilla-La Mancha)
Federated LearningExplainability and InterpretabilityTabularBenchmark
🎯 What it does: Propose a greedy structural consensus algorithm (MCBNC) based on minimum cut analysis to obtain sparse, interpretable consensus DAGs from multiple Bayesian Network (BN) structures in federated learning or multi-model aggregation scenarios.
Bayesian Neural Networks for One-to-Many Mapping in Image Enhancement
Guoxi Huang (University of Bristol), Nantheera Anantrasirichai
RestorationDiffusion modelFlow-based ModelImage
🎯 What it does: Propose the Bayesian Enhancement Model (BEM), which addresses the one-to-many mapping in image enhancement through Bayesian Neural Networks (BNN) and achieves efficient inference via a two-stage BNN-DNN framework.
BayesVQA: Energy-Guided Bayesian Debiasing for Language-Bias-Robust Visual Question Answering
Zhiqi Huang (Beijing Institute of Technology), Bingzhi Chen (Beijing Institute of Technology)
Vision Language ModelContrastive LearningMultimodalityBenchmark
🎯 What it does: Propose an energy-guided Bayesian debiasing framework, BayesVQA, which dynamically regulates prior variance, posterior sampling, and likelihood reweighting through energy uncertainty, and balances visual and language features by correcting perturbations via latent variables;
BCE3S: Binary Cross-Entropy Based Tripartite Synergistic Learning for Long-Tailed Recognition
Weijia Fan (Shenzhen University), Xiaoyang Peng (Sun Yat-sen University)
ClassificationRecognitionContrastive LearningImage
🎯 What it does: Propose a three-branch collaborative learning framework based on binary cross-entropy (BCE3S) to address the long-tailed recognition problem;
BD-Net: Has Depth-Wise Convolution Ever Been Applied in Binary Neural Networks?
DoYoung Kim (Sungkyunkwan University), Jee-Hyong Lee (Sungkyunkwan University)
ClassificationComputational EfficiencyConvolutional Neural NetworkImage
🎯 What it does: Proposed BD-Net, achieving efficient binarization of depthwise separable convolutions in binary neural networks for the first time.
BDD2Seq: Enabling Scalable Reversible-Circuit Synthesis via Graph-to-Sequence Learning
Mingkai Miao (Hong Kong University of Science and Technology), Hongce Zhang (Hong Kong University of Science and Technology)
OptimizationComputational EfficiencyGraph Neural NetworkGraphBenchmark
🎯 What it does: Proposes a graph-to-sequence learning-based BDD variable ordering framework named BDD2Seq for reversible circuit synthesis;
BDLF-Qwen3: Enhanced Cross-Architecture Binary Function Similarity Detection Through Binary Dynamic Layer Fusion
Yuanda Wang (Peking University), Chao Zhang (Tsinghua University)
RetrievalTransformerLarge Language ModelContrastive LearningSequential
🎯 What it does: This study proposes a cross-architecture binary function similarity detection framework named BDLF-Qwen3, and constructs a large-scale cross-architecture binary function dataset named Cross-Bin.
BeDKD: Backdoor Defense Based on Directional Mapping Module and Adversarial Knowledge Distillation
Zhengxian Wu (China Agricultural University), Yiming Xue (China Agricultural University)
ClassificationKnowledge DistillationAdversarial AttackText
🎯 What it does: Propose a backdoor defense framework called BeDKD based on the direction mapping module and adversarial knowledge distillation, which balances defense effectiveness and model performance using a small number of clean samples and contaminated data.