The paper investigates how CLIP-style dual encoders sometimes increase similarity scores when a text description is extended with a plausible but incorrect detail, a phenomenon termed "half-truths." On the COCO dataset, CLIP correctly prefers the shorter, accurate description only 40.6% of the time, dropping to 32.9% when the added detail is a relation. The authors attribute this to weak supervision on caption components and propose CS-CLIP, a component-supervised fine-tuning approach that raises half-truth accuracy to 69.3% and improves compositional benchmark performance by 5.7 points.
By Bora Kargi, Arnas Uselis, Seong Joon Oh
The paper introduces HN-CLIP, a new objective for dense-caption retrieval that adapts similarity margins per negative example using the text encoder’s own geometry. By adding a detached caption‑similarity matrix to the negative logits, HN‑CLIP addresses the issue of near‑duplicate captions that cause premature loss saturation in InfoNCE training. Experiments on four benchmarks show that HN‑CLIP outperforms leading methods by 2.4–4.3 R@1, trains 2.4× faster than GOAL and 5.4× faster than StructXLIP, and achieves the best full‑data baseline with only 20% of the training data.
By Haoyue Liu, Ye Chen, Zhichao Wang, Xiaoying Tang
Flow-CLIP introduces a method to improve compositional image-text matching by refining global image-text alignment with image-conditioned changes in token–patch evidence. It combines Global Compatibility Calibration (GCC) for sentence-level similarity, Image-Conditioned Evidence Refinement (ICER) to incorporate visual patch context, and Cross-Image Margin Ranking (CIMR) to enhance ranking margins when images support correct descriptions. Experiments demonstrate that Flow-CLIP outperforms pretrained CLIP and other global alignment approaches on binding accuracy and reduces reliance on language priors.
By Minghong Wei, Enguang Zhu, Pu Cao, Lu Yang, Qing Song
DEPICT is a new training‑free metric for evaluating text‑to‑image alignment. It replaces fixed reference answers with an agreement rule that compares image‑based and caption‑only responses, weighting questions by how decisively the caption determines them. By merging this agreement score with a holistic score, DEPICT improves negation accuracy dramatically and outperforms existing training‑free metrics while matching or exceeding fine‑tuned evaluators on several benchmarks.
By Vasco Ramos, Sandra Godinho Silva, Joao Magalhaes, Ricardo Rei, Pedro Henrique Martins
arXiv:2607. 28986v1 Announce Type: cross Abstract: Zero-shot image captioning (ZIC) describes images without paired image-caption supervision during captioner training, relying on text-only corpora and frozen pretrained image-text scorers.
By Duy Tran Thanh, Thien-Phuc Doan, Long Nguyen-Vu, Ngo Tan Vu Khanh
The paper introduces Preserve-and-Compose Training (PACT) for composed image retrieval, a task where a query image is modified by a textual instruction while preserving visual content from a reference image. PACT learns from image–text–text triplets, using target captions for supervision and visual evidence from the source image to maintain relevant details, without requiring target images or gallery updates. The authors also propose Chord scoring, which blends target similarity with source-relative directional agreement in a frozen image space, and demonstrate that this combined approach yields strong retrieval performance across multiple zero-shot CIR benchmarks and various backbones.
By Sehyun Kwon
arXiv:2609.23717v1 Announce Type: new
Abstract: Global vision--language similarities compress an image and a caption into one vector, preserving semantics but not which word corresponds to which regi...
By Liuyang Song, Yi Zhang, Zhongyi Deng, Daqian Yang, Hongbo Zhang
The paper introduces Visual Retrieval Heads (VRHs), a small fraction of attention heads in vision‑language models that are causally responsible for grounding text descriptions to image regions. By recasting head‑scoring methods and evaluating across eleven VLMs and five benchmarks, the authors show that masking the top 20 VRHs can drop grounding accuracy by up to 80 percentage points, while random masking has little effect. VRHs generalize across various visual reference tasks, preserve output format while corrupting localization, and transfer causally across models sharing an LLM backbone.
By Chanho Park, Daehyeon Choi, Jihyun Lee, Minhyuk Sung
arXiv:2609.00591v1 Announce Type: new
Abstract: An image may be worth a thousand words, but most captioning models describe it in only a few. Modern vision-language models produce fluent high-level c...
By Suryaansh Jain, Rahasya Barkur, Vishal G, Ryan Rossi, Franck Dernoncourt, Jack Wang, Koustava Goswami, Nedim Lipka, Puneet Mathur, Samyadeep Basu, Seunghyun Yoon
Global vision--language similarities compress an image and a caption into one vector, preserving semantics but not which word corresponds to which region or how those regions are arranged; a model can...
The paper demonstrates that vision‑language models (VLMs) possess a small set of attention heads, called Visual Retrieval Heads (VRHs), that are causally responsible for linking text prompts to specific image regions. By adapting head‑scoring techniques from language models, the authors identify VRHs as the heads whose attention from output prediction tokens, summed over the ground‑truth referent region, most reliably indicates causal grounding. Experiments across eleven VLMs and five referring‑expression benchmarks show that masking the top 20 VRHs can drop grounding accuracy by up to 80 percentage points, while random masking has little effect, and that VRHs generalize across diverse visual tasks and transfer across models sharing an LLM backbone.
The paper presents a matched‑budget audit framework for evaluating recaptioned image‑text supervision distributions, which are created by applying a captioning policy, captioner, and source corpus to generate dense descriptions for text‑to‑image training. The framework operates under a fixed text budget of 64 tokens and produces a five‑axis profile covering prompt coverage, image‑conditioned faithfulness, and caption surface quality, using controllable basic units as a common claim metric. The authors apply the framework to seven paired comparisons across five public corpora, demonstrating improvements in supported claim units and revealing consistent long‑vs‑dense trade‑offs across different judges and budgets, and they release the audited multi‑source recap corpus and audit artifacts.
By Giyeong Oh, Junghun Park, Yuhan Bae, Youngjae Yu