MMEB-V3 benchmark shows omni-modality embedding models fail to enforce instruction-specified modality constraints and exhibit asymmetric, query-biased retrieval.
citation dossier
Gme: Improving universal multimodal retrieval by multimodal llms
17Pith papers citing it
19reference links
cs.CVtop field · 9 papers
UNVERDICTEDtop verdict bucket · 16 papers
why this work matters in Pith
Pith has found this work in 17 reviewed papers. Its strongest current cluster is cs.CV (9 papers). The largest review-status bucket among citing papers is UNVERDICTED (16 papers). For highly cited works, this page shows a dossier first and a bounded explorer second; it never tries to render every citing paper at once.
years
2026 17representative citing papers
CodeMMR creates a unified embedding space for text, code, and images, outperforming baselines by 10 nDCG@10 points and boosting RAG code generation quality.
Bottleneck Tokens paired with a masked generative objective achieve state-of-the-art unified multimodal retrieval performance among 2B-scale models on the MMEB-V2 benchmark with 78 datasets.
ColChunk adaptively chunks visual document patches into contextual multi-vectors via clustering, cutting storage by over 90% while raising average nDCG@5 by 9 points.
MARVEL reaches 37.9 nDCG@10 on the MM-BRIGHT benchmark by combining LLM query expansion, a reasoning-enhanced dense retriever, and GPT-4o CoT reranking, beating prior multimodal encoders by 10.3 points.
Introduces the OACIR task requiring instance-level consistency via bounding-box anchors, a 160K real-world benchmark OACIRR, and the AdaFocal framework that adaptively focuses attention on the anchored region.
PLUME uses latent-state autoregressive rollouts and a progressive training curriculum to deliver efficient reasoning for universal multimodal embeddings without generating explicit rationales.
GELATO extends frozen text embedding models with locked image and audio encoders, training minimal connectors to produce a single semantic embedding space for text, image, audio, and video while keeping original text performance unchanged.
A scalable training-free pipeline using video segmentation, filtering, and off-the-shelf multimodal models creates DenseStep2M, a dataset of 100K videos and 2M detailed instructional steps that improves dense captioning, step grounding, and cross-modal retrieval.
Rewrite-driven generation with alignment and RL produces shorter, more effective generative multimodal embeddings than CoT methods on retrieval benchmarks.
MiMIC mitigates visual modality collapse and semantic misalignment in universal multimodal retrieval via fusion-in-decoder architecture and robust single-modality training.
SLQ adapts frozen MLLMs for multimodal retrieval by appending shared latent queries to text and image tokens and introduces KARR-Bench to test knowledge-aware reasoning retrieval.
ViLL-E introduces a dynamic embedding mechanism and joint contrastive-generative training for VideoLLMs, delivering up to 7% gains in temporal localization and 4% in video retrieval while enabling new zero-shot capabilities.
HIVE raises multimodal retrieval nDCG@10 to 41.7 on the MM-BRIGHT benchmark by inserting LLM-driven hypothesis generation and verification between retrieval passes, delivering +9.5 over the best text-only baseline and +14.1 over the best multimodal baseline.
TriAlignGR integrates visual content and latent user interests into Semantic IDs via cross-modal alignment, CoT-based interest mining, and triangular multitask training to address content degradation and semantic opacity in generative recommenders.
SSA-ME uses saliency-aware modeling to reduce visual neglect and semantic drift, achieving SOTA results on the MMEB benchmark for multimodal retrieval.
BRIDGE reaches 29.7 nDCG@10 on MM-BRIGHT by RL-aligning multimodal queries to text and using a reasoning retriever, beating multimodal encoders and, when combined with Nomic-Vision, exceeding the best text-only retriever at 33.3.
citing papers explorer
-
MMEB-V3: Measuring the Performance Gaps of Omni-Modality Embedding Models
MMEB-V3 benchmark shows omni-modality embedding models fail to enforce instruction-specified modality constraints and exhibit asymmetric, query-biased retrieval.
-
CodeMMR: Bridging Natural Language, Code, and Image for Unified Retrieval
CodeMMR creates a unified embedding space for text, code, and images, outperforming baselines by 10 nDCG@10 points and boosting RAG code generation quality.
-
Bottleneck Tokens for Unified Multimodal Retrieval
Bottleneck Tokens paired with a masked generative objective achieve state-of-the-art unified multimodal retrieval performance among 2B-scale models on the MMEB-V2 benchmark with 78 datasets.
-
Visual Late Chunking: An Empirical Study of Contextual Chunking for Efficient Visual Document Retrieval
ColChunk adaptively chunks visual document patches into contextual multi-vectors via clustering, cutting storage by over 90% while raising average nDCG@5 by 9 points.
-
MARVEL: Multimodal Adaptive Reasoning-intensiVe Expand-rerank and retrievaL
MARVEL reaches 37.9 nDCG@10 on the MM-BRIGHT benchmark by combining LLM query expansion, a reasoning-enhanced dense retriever, and GPT-4o CoT reranking, beating prior multimodal encoders by 10.3 points.
-
Beyond Semantic Search: Towards Referential Anchoring in Composed Image Retrieval
Introduces the OACIR task requiring instance-level consistency via bounding-box anchors, a 160K real-world benchmark OACIRR, and the AdaFocal framework that adaptively focuses attention on the anchored region.
-
PLUME: Latent Reasoning Based Universal Multimodal Embedding
PLUME uses latent-state autoregressive rollouts and a progressive training curriculum to deliver efficient reasoning for universal multimodal embeddings without generating explicit rationales.
-
jina-embeddings-v5-omni: Geometry-preserving Embeddings via Locked Aligned Towers
GELATO extends frozen text embedding models with locked image and audio encoders, training minimal connectors to produce a single semantic embedding space for text, image, audio, and video while keeping original text performance unchanged.
-
DenseStep2M: A Scalable, Training-Free Pipeline for Dense Instructional Video Annotation
A scalable training-free pipeline using video segmentation, filtering, and off-the-shelf multimodal models creates DenseStep2M, a dataset of 100K videos and 2M detailed instructional steps that improves dense captioning, step grounding, and cross-modal retrieval.
-
Beyond Chain-of-Thought: Rewrite as a Universal Interface for Generative Multimodal Embeddings
Rewrite-driven generation with alignment and RL produces shorter, more effective generative multimodal embeddings than CoT methods on retrieval benchmarks.
-
MiMIC: Mitigating Visual Modality Collapse in Universal Multimodal Retrieval While Avoiding Semantic Misalignment
MiMIC mitigates visual modality collapse and semantic misalignment in universal multimodal retrieval via fusion-in-decoder architecture and robust single-modality training.
-
SLQ: Bridging Modalities via Shared Latent Queries for Retrieval with Frozen MLLMs
SLQ adapts frozen MLLMs for multimodal retrieval by appending shared latent queries to text and image tokens and introduces KARR-Bench to test knowledge-aware reasoning retrieval.
-
ViLL-E: Video LLM Embeddings for Retrieval
ViLL-E introduces a dynamic embedding mechanism and joint contrastive-generative training for VideoLLMs, delivering up to 7% gains in temporal localization and 4% in video retrieval while enabling new zero-shot capabilities.
-
HIVE: Query, Hypothesize, Verify An LLM Framework for Multimodal Reasoning-Intensive Retrieval
HIVE raises multimodal retrieval nDCG@10 to 41.7 on the MM-BRIGHT benchmark by inserting LLM-driven hypothesis generation and verification between retrieval passes, delivering +9.5 over the best text-only baseline and +14.1 over the best multimodal baseline.
-
TriAlignGR: Triangular Multitask Alignment with Multimodal Deep Interest Mining for Generative Recommendation
TriAlignGR integrates visual content and latent user interests into Semantic IDs via cross-modal alignment, CoT-based interest mining, and triangular multitask training to address content degradation and semantic opacity in generative recommenders.
-
Combating Visual Neglect and Semantic Drift in Large Multimodal Models for Enhanced Cross-Modal Retrieval
SSA-ME uses saliency-aware modeling to reduce visual neglect and semantic drift, achieving SOTA results on the MMEB benchmark for multimodal retrieval.
-
BRIDGE: Multimodal-to-Text Retrieval via Reinforcement-Learned Query Alignment
BRIDGE reaches 29.7 nDCG@10 on MM-BRIGHT by RL-aligning multimodal queries to text and using a reasoning retriever, beating multimodal encoders and, when combined with Nomic-Vision, exceeding the best text-only retriever at 33.3.