arXiv Machine Learning July 28, 2026 By Sultan Alshehri, Zhantao Yang, Han Zhang, Marios Savvides

Similarity Is Not Logic: Factored Inference for Dual-Encoder Vision-Language Models

Read the original on arXiv Machine Learning →

arXiv:2607. 23052v1 Announce Type: cross Abstract: Dual-encoder vision-language models (VLMs) expose a similarity interface that enables zero-shot retrieval but fails compositional constraints: queries like "umbrella and no person" retrieve images containing both, even when concept detection is reliable.

Summary generated by The Flow from the publisher's feed. The full article lives at arXiv Machine Learning.

arXiv AI

Jun 16

Beyond Scalar Distances: Semantic Attribute Gradients from Frozen MLLMs for Visual Embeddings

arXiv:2606. 15134v1 Announce Type: cross Abstract: Vision encoders for retrieval are typically trained with class-label supervision: each training pair reduces to a scalar that uniformly pushes the embedding apart or pulls it together, as if every visual attribute either differed or matched.

By Shubhang Bhatnagar, Dheeraj Baiju, Narendra Ahuja

llms rag efficiency multimodal benchmarks

arXiv AI

Jul 8

CLAY: Conditional Visual Similarity Modulation in Vision-Language Embedding Space

arXiv:2604. 11539v2 Announce Type: replace-cross Abstract: Human perception of visual similarity is inherently adaptive and subjective, depending on the users' interests and focus.

By Sohwi Lim, Lee Hyoseok, Jungjoon Park, Tae-Hyun Oh

llms rag multimodal

Hugging Face Trending Papers

Jul 6

DiCE-CIR: Direct Composition Learning for Efficient Zero-Shot Composed Image Retrieval

Zero-shot composed image retrieval (ZS-CIR) aims to retrieve a target image from a multimodal query consisting of a reference image and an edit text describing the desired modification. Recent ZS-CIR studies have relied on projection-based methods that map a reference image into pseudo-word tokens in the text embedding space.

llms rag multimodal benchmarks safety

arXiv Machine Learning

Aug 3

FORGE: Frame Orthogonality in Relevance Geometry for Long-Form Video Understanding

arXiv:2607. 25266v2 Announce Type: replace-cross Abstract: Multimodal large language models (MLLMs) have enabled long-form video understanding at a scale that was not previously possible.

By Ghazal Kaviani, Ghassan AlRegib

llms rag nlp multimodal

arXiv AI

Jul 7

Text as Partial Constraint: Core-Residual Alignment for Robust Vision-Language Learning

arXiv:2607. 03143v1 Announce Type: cross Abstract: Vision-language alignment powers open-vocabulary recognition, retrieval, and LVLM grounding, yet natural captions are often underspecified, making similarity brittle and overly confident under paraphrase and omitted details.

By Chengzhen Yu, Canran Xiao, Siyuan Ma, Yang Liu

multimodal safety

arXiv AI

Jul 1

Thinking Before Retrieving: Robust Zero-Shot Composed Image Retrieval via Strategic Planning and Self-Criticism

arXiv:2606. 31222v1 Announce Type: new Abstract: Composed image retrieval requires identifying a target image from a gallery by integrating a reference image with a textual modification instruction.

By Gunho Jung, Jeong-Woo Park, Seon Bin Kim, Seong-Whan Lee

rag

Related stories

Beyond Scalar Distances: Semantic Attribute Gradients from Frozen MLLMs for Visual Embeddings

CLAY: Conditional Visual Similarity Modulation in Vision-Language Embedding Space

DiCE-CIR: Direct Composition Learning for Efficient Zero-Shot Composed Image Retrieval

FORGE: Frame Orthogonality in Relevance Geometry for Long-Form Video Understanding

Text as Partial Constraint: Core-Residual Alignment for Robust Vision-Language Learning

Thinking Before Retrieving: Robust Zero-Shot Composed Image Retrieval via Strategic Planning and Self-Criticism