CiteVQA requires models to cite specific document regions with bounding boxes alongside answers and finds that even the strongest MLLMs frequently cite the wrong region, with top SAA scores of only 76.0 for closed models and 22.5 for open-source ones.
hub Canonical reference
Kosmos-2: Grounding Multimodal Large Language Models to the World
Canonical reference. 83% of citing Pith papers cite this work as background.
82
Pith papers citing it
Background
83% of classified citations
abstract
We introduce Kosmos-2, a Multimodal Large Language Model (MLLM), enabling new capabilities of perceiving object descriptions (e.g., bounding boxes) and grounding text to the visual world. Specifically, we represent refer expressions as links in Markdown, i.e., ``[text span](bounding boxes)'', where object descriptions are sequences of location tokens. Together with multimodal corpora, we construct large-scale data of grounded image-text pairs (called GrIT) to train the model. In addition to the existing capabilities of MLLMs (e.g., perceiving general modalities, following instructions, and performing in-context learning), Kosmos-2 integrates the grounding capability into downstream applications. We evaluate Kosmos-2 on a wide range of tasks, including (i) multimodal grounding, such as referring expression comprehension, and phrase grounding, (ii) multimodal referring, such as referring expression generation, (iii) perception-language tasks, and (iv) language understanding and generation. This work lays out the foundation for the development of Embodiment AI and sheds light on the big convergence of language, multimodal perception, action, and world modeling, which is a key step toward artificial general intelligence. Code and pretrained models are available at https://aka.ms/kosmos-2.
hub tools
citation-role summary
background 28
dataset 5
method 2
citation-polarity summary
claims ledger
- abstract We introduce Kosmos-2, a Multimodal Large Language Model (MLLM), enabling new capabilities of perceiving object descriptions (e.g., bounding boxes) and grounding text to the visual world. Specifically, we represent refer expressions as links in Markdown, i.e., ``[text span](bounding boxes)'', where object descriptions are sequences of location tokens. Together with multimodal corpora, we construct large-scale data of grounded image-text pairs (called GrIT) to train the model. In addition to the existing capabilities of MLLMs (e.g., perceiving general modalities, following instructions, and per
co-cited works
representative citing papers
Molmo VLMs trained on newly collected PixMo open datasets achieve state-of-the-art performance among open-weight models and surpass multiple proprietary VLMs including Claude 3.5 Sonnet and Gemini 1.5 Pro.
MMMU provides 11.5K heterogeneous college-level multimodal questions that current models solve at 56-59% accuracy, establishing a new standard for expert multimodal evaluation.
HEED replaces uniform residual alignment with density-weighted alignment using patch self-dissimilarity to improve hybrid VLM distillation, gaining 8.7 points on OCRBench v2 and 5.13 on a 10-benchmark average.
INSET embeds images as native tokens in interleaved instructions, outperforming prior methods on multi-image consistency and text alignment as complexity grows.
RevealLayer decomposes natural images into multiple RGBA layers using diffusion models with region-aware attention, occlusion-guided adaptation, and a composite loss, outperforming prior methods on a new benchmark dataset.
GazeVLM introduces internal gaze tokens that allow VLMs to dynamically suppress irrelevant visual features and simulate foveal attention for improved high-resolution multimodal reasoning.
Introduces closed-set C-Bench and open-set O-Bench for layout-guided diffusion models, a unified semantic-spatial scoring protocol, and ranks six models after generating and evaluating 319,086 images.
MMR-AD is a new benchmark dataset showing that current generalist MLLMs lag industrial needs for anomaly detection, with Anomaly-R1 delivering better results through reasoning and RL.
STORM is an end-to-end MLLM for referring multi-object tracking that uses task-composition learning to leverage sub-task data and introduces the STORM-Bench dataset, achieving SOTA results.
MARINER is a new benchmark dataset and evaluation framework for fine-grained perception and causal reasoning in open-water scenes using 16,629 images across 63 vessel categories, diverse environments, and maritime incidents.
Mema adds a stateful memory module to vision encoders that accumulates hierarchical visual features across layers and selectively injects portions back via feedback to preserve fine-grained cues, yielding consistent gains on multimodal benchmarks.
OR-VSKC provides 28,190 synthetic operating room images plus an expert subset to expose and reduce visual-semantic knowledge conflicts in multimodal models for surgical risk detection.
A survey that defines Compound AI Systems, proposes a multi-dimensional taxonomy based on component roles and orchestration strategies, reviews four foundational paradigms, and identifies key challenges for future research.
CompART adds a composition loss on decomposed captions to regularize attention sums and improves multi-object grounding plus VQA across four VLM types and six benchmarks.
Scaled vanilla autoregressive models based on Llama achieve 2.18 FID on ImageNet 256x256 image generation, beating popular diffusion models without visual inductive biases.
3D-VLA is a new embodied foundation model that uses a 3D LLM plus aligned diffusion models to generate future images and point clouds for improved reasoning and action planning in 3D environments.
SEED-Bench is a new benchmark of 19K multiple-choice questions for evaluating generative comprehension in multimodal LLMs across 12 image and video dimensions.
PRISM benchmark perturbs Crello layouts into 110K samples isolating design principle violations, reveals limited sensitivity in several multimodal models, and proposes a multi-scale framework combining scorers, instruction-tuned VLMs, and prompt methods for interpretable design assessment.
WOW-Seg proposes a word-free open-world segmentation model using Mask2Token and Cascade Attention Mask modules, reporting 89.7 semantic similarity and 82.4 semantic IoU on LVIS with one-eighth the parameters of prior SOTA plus a new 7,662-class benchmark.
Controlla learns identity and attribute factors from multimodal inputs and aligns them with graph priors using graph-constrained optimal transport to enforce consistent attribute trajectories while preserving reference identity.
Controlled counterfactual perturbations reveal no correlation between embedding cosine similarity and approximation behavior in two visual grounding models.
LLaVA-UHD v4 reduces visual-encoding FLOPs by 55.8% for high-resolution images in MLLMs via slice-based encoding plus intra-ViT early compression while matching or exceeding baseline performance on document, OCR, and VQA benchmarks.
AlbumFill retrieves identity-consistent references from personal albums via VLM-inferred semantic cues to support personalized image completion.
citing papers explorer
-
CiteVQA: Benchmarking Evidence Attribution for Trustworthy Document Intelligence
CiteVQA requires models to cite specific document regions with bounding boxes alongside answers and finds that even the strongest MLLMs frequently cite the wrong region, with top SAA scores of only 76.0 for closed models and 22.5 for open-source ones.
-
Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models
Molmo VLMs trained on newly collected PixMo open datasets achieve state-of-the-art performance among open-weight models and surpass multiple proprietary VLMs including Claude 3.5 Sonnet and Gemini 1.5 Pro.
-
MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI
MMMU provides 11.5K heterogeneous college-level multimodal questions that current models solve at 56-59% accuracy, establishing a new standard for expert multimodal evaluation.
-
HEED: Density-Weighted Residual Alignment for Hybrid Vision-Language Model Distillation
HEED replaces uniform residual alignment with density-weighted alignment using patch self-dissimilarity to improve hybrid VLM distillation, gaining 8.7 points on OCRBench v2 and 5.13 on a 10-benchmark average.
-
Images in Sentences: Scaling Interleaved Instructions for Unified Visual Generation
INSET embeds images as native tokens in interleaved instructions, outperforming prior methods on multi-image consistency and text alignment as complexity grows.
-
RevealLayer: Disentangling Hidden and Visible Layers via Occlusion-Aware Image Decomposition
RevealLayer decomposes natural images into multiple RGBA layers using diffusion models with region-aware attention, occlusion-guided adaptation, and a composite loss, outperforming prior methods on a new benchmark dataset.
-
GazeVLM: Active Vision via Internal Attention Control for Multimodal Reasoning
GazeVLM introduces internal gaze tokens that allow VLMs to dynamically suppress irrelevant visual features and simulate foveal attention for improved high-resolution multimodal reasoning.
-
Benchmarking Layout-Guided Diffusion Models through Unified Semantic-Spatial Evaluation in Closed and Open Settings
Introduces closed-set C-Bench and open-set O-Bench for layout-guided diffusion models, a unified semantic-spatial scoring protocol, and ranks six models after generating and evaluating 319,086 images.
-
MMR-AD: A Large-Scale Multimodal Dataset for Benchmarking General Anomaly Detection with Multimodal Large Language Models
MMR-AD is a new benchmark dataset showing that current generalist MLLMs lag industrial needs for anomaly detection, with Anomaly-R1 delivering better results through reasoning and RL.
-
STORM: End-to-End Referring Multi-Object Tracking in Videos
STORM is an end-to-end MLLM for referring multi-object tracking that uses task-composition learning to leverage sub-task data and introduces the STORM-Bench dataset, achieving SOTA results.
-
MARINER: A 3E-Driven Benchmark for Fine-Grained Perception and Complex Reasoning in Open-Water Environments
MARINER is a new benchmark dataset and evaluation framework for fine-grained perception and causal reasoning in open-water scenes using 16,629 images across 63 vessel categories, diverse environments, and maritime incidents.
-
Mema: Memory-Augmented Adapter for Enhanced Vision-Language Understanding
Mema adds a stateful memory module to vision encoders that accumulates hierarchical visual features across layers and selectively injects portions back via feedback to preserve fine-grained cues, yielding consistent gains on multimodal benchmarks.
-
OR-VSKC: Resolving Visual-Semantic Knowledge Conflicts in Operating Rooms with Synthetic Data-Guided Alignment
OR-VSKC provides 28,190 synthetic operating room images plus an expert subset to expose and reduce visual-semantic knowledge conflicts in multimodal models for surgical risk detection.
-
From Standalone LLMs to Integrated Intelligence: A Survey of Compound Al Systems
A survey that defines Compound AI Systems, proposes a multi-dimensional taxonomy based on component roles and orchestration strategies, reviews four foundational paradigms, and identifies key challenges for future research.
-
The ART of Composition: Attention-Regularized Training for Compositional Visual Grounding
CompART adds a composition loss on decomposed captions to regularize attention sums and improves multi-object grounding plus VQA across four VLM types and six benchmarks.
-
Autoregressive Model Beats Diffusion: Llama for Scalable Image Generation
Scaled vanilla autoregressive models based on Llama achieve 2.18 FID on ImageNet 256x256 image generation, beating popular diffusion models without visual inductive biases.
-
3D-VLA: A 3D Vision-Language-Action Generative World Model
3D-VLA is a new embodied foundation model that uses a 3D LLM plus aligned diffusion models to generate future images and point clouds for improved reasoning and action planning in 3D environments.
-
SEED-Bench: Benchmarking Multimodal LLMs with Generative Comprehension
SEED-Bench is a new benchmark of 19K multiple-choice questions for evaluating generative comprehension in multimodal LLMs across 12 image and video dimensions.
-
Through the PRISM: Principle-Aware, Interpretable, and Multi-Scale Evaluation of Visual Designs
PRISM benchmark perturbs Crello layouts into 110K samples isolating design principle violations, reveals limited sensitivity in several multimodal models, and proposes a multi-scale framework combining scorers, instruction-tuned VLMs, and prompt methods for interpretable design assessment.
-
WOW-Seg: A Word-free Open World Segmentation Model
WOW-Seg proposes a word-free open-world segmentation model using Mask2Token and Cascade Attention Mask modules, reporting 89.7 semantic similarity and 82.4 semantic IoU on LVIS with one-eighth the parameters of prior SOTA plus a new 7,662-class benchmark.
-
Controlla: Learning Controllability via Graph-Constrained Latent Geometry
Controlla learns identity and attribute factors from multimodal inputs and aligns them with graph priors using graph-constrained optimal transport to enforce consistent attribute trajectories while preserving reference identity.
-
Investigating Anisotropy in Visual Grounding under Controlled Counterfactual Perturbations
Controlled counterfactual perturbations reveal no correlation between embedding cosine similarity and approximation behavior in two visual grounding models.
-
LLaVA-UHD v4: What Makes Efficient Visual Encoding in MLLMs?
LLaVA-UHD v4 reduces visual-encoding FLOPs by 55.8% for high-resolution images in MLLMs via slice-based encoding plus intra-ViT early compression while matching or exceeding baseline performance on document, OCR, and VQA benchmarks.
-
AlbumFill: Album-Guided Reasoning and Retrieval for Personalized Image Completion
AlbumFill retrieves identity-consistent references from personal albums via VLM-inferred semantic cues to support personalized image completion.
-
CoNewsReader: Supporting Comprehensive Understanding and Raising Critical Thoughts on Social Media News Through Comments
CoNewsReader integrates user comments with an LLM to improve critical news reading on social media, with a 24-participant study showing gains in comprehension and critical thinking over baseline interfaces.
-
Latent Denoising Improves Visual Alignment in Large Multimodal Models
A latent denoising objective with saliency-aware corruption and contrastive distillation improves visual alignment and corruption robustness in large multimodal models.
-
G-MIXER: Geodesic Mixup-based Implicit Semantic Expansion and Explicit Semantic Re-ranking for Zero-Shot Composed Image Retrieval
G-MIXER achieves state-of-the-art zero-shot composed image retrieval by using geodesic mixup to build diverse implicit candidates and MLLM-derived explicit semantics for re-ranking.
-
CoME-VL: Scaling Complementary Multi-Encoder Vision-Language Learning
CoME-VL fuses contrastive and self-supervised vision encoders via entropy-guided multi-layer aggregation and RoPE cross-attention to improve vision-language model performance on benchmarks.
-
MapTab: Are MLLMs Ready for Multi-Criteria Route Planning in Heterogeneous Graphs?
MapTab is a new multimodal benchmark with 328 images and nearly 200k queries that shows current MLLMs have substantial difficulty with multi-criteria route planning when visual and tabular information must be combined.
-
Memento: Towards Proactive Visualization of Everyday Memories with Personal Wearable AR Assistant
Memento captures verbal queries with spatiotemporal contexts, discovers recurring interest patterns, and proactively delivers updated AR responses when matching situations are detected.
-
OmniDrive-R1: Reinforcement-driven Interleaved Multi-modal Chain-of-Thought for Trustworthy Vision-Language Autonomous Driving
OmniDrive-R1 boosts VLM reasoning score from 51.77% to 80.35% and answer accuracy from 37.81% to 73.62% on DriveLMM-o1 via reinforcement-driven interleaved multi-modal chain-of-thought with annotation-free grounding.
-
DeFacto: Counterfactual Thinking with Images for Enforcing Evidence-Grounded and Faithful Reasoning
DeFacto trains multimodal models with counterfactual image variants and GRPO reinforcement learning to enforce that correct answers are supported by correct visual evidence.
-
Training-Free Multimodal Large Language Model Orchestration
LLM Orchestration integrates modality experts via an LLM controller, cross-modal memory, and interaction layer to enable multimodal input-output without gradient-based training.
-
GraspVLA: a Grasping Foundation Model Pre-trained on Billion-scale Synthetic Action Data
GraspVLA shows that pretraining a grasping model on a billion synthetic action frames enables zero-shot open-vocabulary performance and sim-to-real transfer.
-
InfiGUI-R1: Advancing Multimodal GUI Agents from Reactive Actors to Deliberative Reasoners
InfiGUI-R1 uses Reasoning Injection via spatial distillation followed by Deliberation Enhancement via RL to evolve GUI agents from reactive actors to deliberative reasoners, reporting strong performance on grounding and trajectory tasks.
-
Perception Encoder: The best visual embeddings are not at the output of the network
Intermediate layers of a contrastively trained vision-language encoder yield stronger general embeddings than the output layer, enabling state-of-the-art performance across image/video classification, multimodal QA, and dense prediction after simple alignment.
-
LMM-R1: Empowering 3B LMMs with Strong Reasoning Abilities Through Two-Stage Rule-Based RL
A two-stage RL framework first boosts text reasoning in 3B LMMs then adapts it to multimodal inputs, producing modest benchmark gains of 4.5-4.8%.
-
Sa2VA: Marrying SAM2 with LLaVA for Dense Grounded Understanding of Images and Videos
Sa2VA unifies SAM-2 segmentation with MLLM reasoning into a single model for referring segmentation and conversation on images and videos, supported by a new 72k-expression Ref-SAV dataset.
-
Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling
InternVL 2.5 is the first open-source MLLM to surpass 70% on the MMMU benchmark via model, data, and test-time scaling, with a 3.7-point gain from chain-of-thought reasoning.
-
LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding
LongVU adaptively compresses long video tokens using DINOv2-based frame deduplication, text-guided cross-modal selection, and temporal spatial reduction to improve video-language understanding in MLLMs with minimal detail loss.
-
SEED-X: Multimodal Models with Unified Multi-granularity Comprehension and Generation
SEED-X is a unified multimodal foundation model that handles multi-granularity visual semantics for both comprehension and generation across arbitrary image sizes and ratios.
-
MM1: Methods, Analysis & Insights from Multimodal LLM Pre-training
MM1 models achieve state-of-the-art few-shot multimodal results by pre-training on a careful mix of image-caption, interleaved, and text-only data with optimized image encoders.
-
MoE-LLaVA: Mixture of Experts for Large Vision-Language Models
MoE-LLaVA applies mixture-of-experts sparsity to LVLMs via MoE-Tuning, delivering LLaVA-1.5-7B level visual understanding and better hallucination resistance with only ~3B active parameters.
-
SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agents
SeeClick improves visual GUI agents via GUI grounding pre-training on automatically curated data and introduces the ScreenSpot benchmark, with results indicating that stronger grounding boosts downstream task performance.
-
GPT-4V(ision) is a Generalist Web Agent, if Grounded
GPT-4V achieves 51.1% success on live web tasks as a generalist agent when plans are manually grounded, outperforming text-only models, but automatic grounding lags far behind oracle performance.
-
ShareGPT4V: Improving Large Multi-Modal Models with Better Captions
A new 1.2M-caption dataset generated via GPT-4V improves LMMs on MME and MMBench by 222.8/22.0/22.3 and 2.7/1.3/1.5 points respectively when used for supervised fine-tuning.
-
Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models
Qwen-Audio trains a unified model on diverse audio and tasks with hierarchical tags to enable strong zero-shot performance on audio understanding benchmarks and multi-turn audio chat.
-
CogVLM: Visual Expert for Pretrained Language Models
CogVLM adds a trainable visual expert inside frozen language model layers for deep vision-language fusion and reports state-of-the-art results on ten cross-modal benchmarks while preserving NLP performance.
-
Retentive Network: A Successor to Transformer for Large Language Models
RetNet is a new sequence modeling architecture that delivers parallel training, constant-time inference, and competitive language modeling performance as a potential replacement for Transformers.
-
Archon: A Unified Multimodal Model for Holistic Digital Human Generation
Archon unifies seven modalities via modality-specific tokenizers and an autoregressive backbone pretrained on 72 tasks, plus a 4x-efficient video reparameterization and stepwise 'Thinking in Modality' procedure, and reports superior or comparable results on digital-human tasks.