AAAI 2026 accepted papers on Vision-language and multimodal
543 of the 4,417 papers accepted at AAAI 2026 (12.3%) are on Vision-language and multimodal, compared with 8.5% at AAAI 2025. They fall into 6 topics.
See AAAI 2026 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Multimodal LLMs · 218 papers
- MedEyes: Learning Dynamic Visual Focus for Medical Progressive Diagnosis (7 citations)
- Toward Multimodal Fake News Detection by Multi-perspective Rationale Generation and Verification (6 citations)
- FaceShield: Explainable Face Anti-Spoofing with Multimodal Large Language Models (4 citations)
- GMAI-VL & GMAI-VL-5.5M: A Large Vision-Language Model and a Comprehensive Multimodal Dataset Towards General Medical AI (4 citations)
- Patho-AgenticRAG: Towards Multimodal Agentic Retrieval-Augmented Generation for Pathology VLMs via Reinforcement Learning (4 citations)
- RemoteReasoner: Towards Unifying Geospatial Reasoning Workflow (4 citations)
- LENS: Learning to Segment Anything with Unified Reinforced Reasoning (3 citations)
- MedAtlas: Evaluating LLMs for Multi-Round, Multi-Task Medical Reasoning Across Diverse Imaging Modalities and Clinical Text (3 citations)
- Multimodal DeepResearcher: Generating Text-Chart Interleaved Reports from Scratch with Agentic Framework (3 citations)
- Patho-R1: A Multimodal Reinforcement Learning-Based Pathology Expert Reasoner (3 citations)
- ReCAD: Reinforcement Learning Enhanced Parametric CAD Model Generation with Vision-Language Models (3 citations)
- Refine-IQA: Multi-Stage Reinforcement Finetuning for Perceptual Image Quality Assessment (3 citations)
and 206 more in the interactive map.
Cross-modal learning · 153 papers
- Hierarchical Prompt Learning for Image- and Text-Based Person Re-Identification (10 citations)
- Text-Guided Channel Perturbation and Pre-Trained Knowledge Integration for Unified Multi-Modality Image Fusion (7 citations)
- MCMoE: Completing Missing Modalities with Mixture of Experts for Incomplete Multimodal Action Quality Assessment (6 citations)
- Distilling Cross-Modal Knowledge via Feature Disentanglement (5 citations)
- EchoMimicV3: 1.3B Parameters Are All You Need for Unified Multi-Modal and Multi-Task Human Animation (5 citations)
- HFR-MKGC: Hierarchical Fusion Reasoning with MLLMs for Multi-modal Knowledge Graph Completion (5 citations)
- Ambiguity-Tolerant Cross-Modal Hashing with Partial Labels (4 citations)
- CADTrack: Learning Contextual Aggregation with Deformable Alignment for Robust RGBT Tracking (4 citations)
- Capturing Dynamic User Interests Under Modality Imbalance for Multimodal Sequential Recommendation (4 citations)
- DGKAN: Dual-branch Graph Kolmogorov-Arnold Network for Unsupervised Multimodal Change Detection (4 citations)
- INTENT: Invariance and Discrimination-aware Noise Mitigation for Robust Composed Image Retrieval (4 citations)
- MdaIF: Robust One-Stop Multi-Degradation-Aware Image Fusion with Language-Driven Semantics (4 citations)
and 141 more in the interactive map.
CLIP and zero-shot · 55 papers
- AdaptCLIP: Adapting CLIP for Universal Visual Anomaly Detection (16 citations)
- RSVG-ZeroOV: Exploring a Training-Free Framework for Zero-Shot Open-Vocabulary Visual Grounding in Remote Sensing Images (5 citations)
- Anomagic: Crossmodal Prompt-driven Zero-shot Anomaly Generation (3 citations)
- CLIPPan: Adapting CLIP as a Supervisor for Unsupervised Pansharpening (3 citations)
- HiMo-CLIP: Modeling Semantic Hierarchy and Monotonicity in Vision-Language Alignment (3 citations)
- InfoCLIP: Bridging Vision-Language Pretraining and Open-Vocabulary Semantic Segmentation via Information-Theoretic Alignment Transfer (3 citations)
- LLM2CLIP: Powerful Language Model Unlocks Richer Cross-Modality Representation (3 citations)
- Plug-and-Play Clarifier: A Zero-Shot Multimodal Framework for Egocentric Intent Disambiguation (2 citations)
- PromptMoE: Generalizable Zero-Shot Anomaly Detection via Visually-Guided Prompt Mixtures (2 citations)
- TIM++: Transductive Information Maximization for Few-Shot CLIP (2 citations)
- Weakly-Supervised Image Forgery Localization via Vision-Language Collaborative Reasoning Framework (2 citations)
- Beyond Retraining: Training-Free Unknown Class Filtering for Source-Free Open Set Domain Adaptation of Vision–Language Models (1 citation)
and 43 more in the interactive map.
Robot action and navigation · 50 papers
- OpenDriveVLA: Towards End-to-end Autonomous Driving with Large Vision Language Action Model (26 citations)
- DexGraspVLA: A Vision-Language-Action Framework Towards General Dexterous Grasping (9 citations)
- VLA-Adapter: An Effective Paradigm for Tiny-Scale Vision-Language-Action Model (9 citations)
- MoLe-VLA: Dynamic Layer-skipping Vision Language Action Model via Mixture-of-Layers for Efficient Robot Manipulation (7 citations)
- SemanticVLA: Semantic-Aligned Sparsification and Enhancement for Efficient Robotic Manipulation (4 citations)
- ReconVLA: Reconstructive Vision-Language-Action Model as Effective Robot Perceiver (3 citations)
- Towards Efficient and Robust Manipulation via Multi-Frame Vision-Language-Action Modeling (3 citations)
- FastDriveVLA: Efficient End-to-End Driving via Plug-and-Play Reconstruction-based Token Pruning (2 citations)
- GraphCoT-VLA: A 3D Spatial-Aware Reasoning Vision-Language-Action Model for Robotic Manipulation with Ambiguous Instructions (2 citations)
- History-Enhanced Two-Stage Transformer for Aerial Vision-and-Language Navigation (2 citations)
- IndoorUAV: Benchmarking Vision-Language UAV Navigation in Continuous Indoor Environments (2 citations)
- Intention-Guided Cognitive Reasoning for Egocentric Long-Term Action Anticipation (2 citations)
and 38 more in the interactive map.
Human and embodied modeling · 46 papers
- HABIT: Chrono-Synergia Robust Progressive Learning Framework for Composed Image Retrieval (5 citations)
- PET2Rep: Towards Vision-Language Model-Drived Automated Radiology Report Generation for Positron Emission Tomography (4 citations)
- PriorRG: Prior-Guided Contrastive Pre-training and Coarse-to-Fine Decoding for Chest X-ray Report Generation (4 citations)
- Beyond Cosine Similarity: Magnitude-Aware CLIP for No-Reference Image Quality Assessment (3 citations)
- Rethinking Saliency Maps: A Cognitive Human Aligned Taxonomy and Evaluation Framework for Explanations (3 citations)
- Cross Modal Fine-grained Alignment via Granularity-aware and Region-uncertain Modeling (2 citations)
- Endowing Vision-Language Models with System 2 Thinking for Fine-grained Visual Recognition (2 citations)
- Heterogeneous Uncertainty-Guided Composed Image Retrieval with Fine-Grained Probabilistic Learning (2 citations)
- KPDM: Key Phrase Dynamic Masking for Robust Text-to-Image Person Retrieval (2 citations)
- Partially Shared Concept Bottleneck Models (2 citations)
- A Disease-Aware Dual-Stage Framework for Chest X-ray Report Generation (1 citation)
- Aware Distillation for Robust Vision-Language Tracking Under Linguistic Sparsity (1 citation)
and 34 more in the interactive map.
Visual question answering · 21 papers
- STRIDE-QA: Visual Question Answering Dataset for Spatiotemporal Reasoning in Urban Driving Scenes (5 citations)
- Cross-Granularity Hypergraph Retrieval-Augmented Generation for Multi-hop Question Answering (3 citations)
- MacVQA: Adaptive Memory Allocation and Global Noise Filtering for Continual Visual Question Answering (2 citations)
- RegionRAG: Region-level Retrieval-Augmented Generation for Visual Document Understanding (2 citations)
- RoadSceneVQA: Benchmarking Visual Question Answering in Roadside Perception Systems for Intelligent Transportation System (2 citations)
- CMID: Towards Medical Visual Question Answering via Contrastive Mutual Information Decoding (1 citation)
- ImageSet2Text: Describing Sets of Images Through Text (1 citation)
- MAVIS: A Benchmark for Multimodal Source Attribution in Long-form Visual Question Answering (1 citation)
- BayesVQA: Energy-Guided Bayesian Debiasing for Language-Bias-Robust Visual Question Answering
- Benchmarking Visual LLMs Resilience to Unanswerable Questions on Visually Rich Documents
- Beyond Static: Related Questions Retrieval Through Conversations in Community Question Answering
- Collaborative Enhancement of Large and Small Models for Question Answering via Dual Knowledge Transfer
and 9 more in the interactive map.