ICLR 2025 accepted papers on Vision-language and multimodal
269 of the 3,703 papers accepted at ICLR 2025 (7.3%) are on Vision-language and multimodal, compared with 5.1% at ICLR 2024. They fall into 6 topics.
See ICLR 2025 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 · 132 papers
- MMed-RAG: Versatile Multimodal RAG System for Medical Vision Language Models (9 citations)
- MedTrinity-25M: A Large-scale Multimodal Dataset with Multigranular Annotations for Medicine (8 citations)
- mPLUG-Owl3: Towards Long Image-Sequence Understanding in Multi-Modal Large Language Models (6 citations)
- G-LLaVA: Solving Geometric Problem with Multi-Modal Large Language Model (5 citations)
- Visual Agents as Fast and Slow Thinkers (4 citations)
- ChartMimic: Evaluating LMM's Cross-Modal Reasoning Capability via Chart-to-Code Generation (3 citations)
- Feast Your Eyes: Mixture-of-Resolution Adaptation for Multimodal Large Language Models (3 citations)
- Mitigating Object Hallucination in MLLMs via Data-augmented Phrase-level Alignment (3 citations)
- MLLM can see? Dynamic Correction Decoding for Hallucination Mitigation (3 citations)
- OBI-Bench: Can LMMs Aid in Study of Ancient Script on Oracle Bones? (3 citations)
- PhysBench: Benchmarking and Enhancing Vision-Language Models for Physical World Understanding (3 citations)
- See It from My Perspective: How Language Affects Cultural Bias in Image Understanding (3 citations)
and 120 more in the interactive map.
CLIP and zero-shot · 40 papers
- MMTEB: Massive Multilingual Text Embedding Benchmark (16 citations)
- CLIBD: Bridging Vision and Genomics for Biodiversity Monitoring at Scale (3 citations)
- EffoVPR: Effective Foundation Model Utilization for Visual Place Recognition (3 citations)
- PathGen-1.6M: 1.6 Million Pathology Image-text Pairs Generation through Multi-agent Collaboration (3 citations)
- Large-scale and Fine-grained Vision-language Pre-training for Enhanced CT Image Understanding (2 citations)
- Repetition Improves Language Model Embeddings (2 citations)
- Diffusion Feedback Helps CLIP See Better (1 citation)
- Efficient and Context-Aware Label Propagation for Zero-/Few-Shot Training-Free Adaptation of Vision-Language Model (1 citation)
- Two Effects, One Trigger: On the Modality Gap, Object Bias, and Information Imbalance in Contrastive Vision-Language Models (1 citation)
- A Simple Framework for Open-Vocabulary Zero-Shot Segmentation
- Attribute-based Visual Reprogramming for Vision-Language Models
- Black Sheep in the Herd: Playing with Spuriously Correlated Attributes for Vision-Language Recognition
and 28 more in the interactive map.
Human and embodied modeling · 37 papers
- The OMG dataset: An Open MetaGenomic corpus for mixed-modality genomic language modeling (44 citations)
- In vivo cell-type and brain region classification via multimodal contrastive learning (11 citations)
- Modeling dynamic social vision highlights gaps between deep learning and humans (10 citations)
- ColPali: Efficient Document Retrieval with Vision Language Models (6 citations)
- Transfusion: Predict the Next Token and Diffuse Images with One Multi-Modal Model (6 citations)
- Scaling Wearable Foundation Models (5 citations)
- The Semantic Hub Hypothesis: Language Models Share Semantic Representations Across Languages and Modalities (5 citations)
- 3D-MolT5: Leveraging Discrete Structural Information for Molecule-Text Modeling (4 citations)
- Finding Shared Decodable Concepts and their Negations in the Brain (1 citation)
- LucidPPN: Unambiguous Prototypical Parts Network for User-centric Interpretable Computer Vision (1 citation)
- OmniCorpus: A Unified Multimodal Corpus of 10 Billion-Level Images Interleaved with Text (1 citation)
- TopoNets: High performing vision and language models with brain-like topography (1 citation)
and 25 more in the interactive map.
Robot action and navigation · 25 papers
- RDT-1B: a Diffusion Foundation Model for Bimanual Manipulation (5 citations)
- AHA: A Vision-Language-Model for Detecting and Reasoning Over Failures in Robotic Manipulation (4 citations)
- Data Scaling Laws in Imitation Learning for Robotic Manipulation (3 citations)
- Towards Realistic UAV Vision-Language Navigation: Platform, Benchmark, and Methodology (3 citations)
- ImagineNav: Prompting Vision-Language Models as Embodied Navigator through Scene Imagination (2 citations)
- LLaRA: Supercharging Robot Learning Data for Vision-Language Policy (2 citations)
- Articulate-Anything: Automatic Modeling of Articulated Objects via a Vision-Language Foundation Model (1 citation)
- HAMSTER: Hierarchical Action Models for Open-World Robot Manipulation (1 citation)
- Robots Pre-train Robots: Manipulation-Centric Robotic Representation from Large-Scale Robot Datasets (1 citation)
- STRAP: Robot Sub-Trajectory Retrieval for Augmented Policy Learning (1 citation)
- TraceVLA: Visual Trace Prompting Enhances Spatial-Temporal Awareness for Generalist Robotic Policies (1 citation)
- Bootstrapping Language-Guided Navigation Learning with Self-Refining Data Flywheel
and 13 more in the interactive map.
Cross-modal learning · 24 papers
- What to align in multimodal contrastive learning? (3 citations)
- Learnable Expansion of Graph Operators for Multi-Modal Feature Fusion (2 citations)
- AnyTouch: Learning Unified Static-Dynamic Representation across Multiple Visuo-tactile Sensors (1 citation)
- Multimodal Lego: Model Merging and Fine-Tuning Across Topologies and Modalities in Biomedicine (1 citation)
- Multiple Heads are Better than One: Mixture of Modality Knowledge Experts for Entity Representation Learning (1 citation)
- OmniBind: Large-scale Omni Multimodal Representation via Binding Spaces (1 citation)
- Test-time Adaptation for Cross-modal Retrieval with Query Shift (1 citation)
- X-Fi: A Modality-Invariant Foundation Model for Multimodal Human Sensing (1 citation)
- An Information Criterion for Controlled Disentanglement of Multimodal Data
- Cross the Gap: Exposing the Intra-modal Misalignment in CLIP via Modality Inversion
- CSA: Data-efficient Mapping of Unimodal Features to Multimodal Features
- EMMA: Empowering Multi-modal Mamba with Structural and Hierarchical Alignment
and 12 more in the interactive map.
Visual question answering · 11 papers
- TEOChat: A Large Vision-Language Assistant for Temporal Earth Observation Data (4 citations)
- Benchmarking Multimodal Retrieval Augmented Generation with Dynamic VQA Dataset and Self-adaptive Planning Agent (3 citations)
- Answer, Assemble, Ace: Understanding How LMs Answer Multiple Choice Questions (1 citation)
- Breaking Neural Network Scaling Laws with Modularity (1 citation)
- Exploring the Effectiveness of Object-Centric Representations in Visual Question Answering: Comparative Insights with Foundation Models (1 citation)
- MLLMs Know Where to Look: Training-free Perception of Small Visual Details with Multimodal LLMs (1 citation)
- MRAG-Bench: Vision-Centric Evaluation for Retrieval-Augmented Multimodal Models (1 citation)
- SV-RAG: LoRA-Contextualizing Adaptation of MLLMs for Long Document Understanding (1 citation)
- Visual Haystacks: A Vision-Centric Needle-In-A-Haystack Benchmark (1 citation)
- VL-ICL Bench: The Devil in the Details of Multimodal In-Context Learning (1 citation)
- CertainlyUncertain: A Benchmark and Metric for Multimodal Epistemic and Aleatoric Awareness