# Do Composed Image Retrieval Benchmarks Require Multimodal Composition?

arXiv: https://arxiv.org/abs/2605.14787 · HTML: https://mlsanity.com/paper/2605.14787/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 37 of its 42 references that are in the map.
- First posted on arXiv 14 May 2026.

## Builds on (most cited first, 30 of 37)

- [Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering](https://mlsanity.com/paper/1612.00837/) (2016)
- [Shortcut Learning in Deep Neural Networks](https://mlsanity.com/paper/2004.07780/) (2020)
- [Annotation Artifacts in Natural Language Inference Data](https://mlsanity.com/paper/1803.02324/) (2018)
- [When and why vision-language models behave like bags-of-words, and what to do about it?](https://mlsanity.com/paper/2210.01936/) (2022)
- [Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality](https://mlsanity.com/paper/2204.03162/) (2022)
- [Don't Just Assume; Look and Answer: Overcoming Priors for Visual Question Answering](https://mlsanity.com/paper/1712.00377/) (2017)
- [Hypothesis Only Baselines in Natural Language Inference](https://mlsanity.com/paper/1805.01042/) (2018)
- [Don't Take the Easy Way Out: Ensemble Based Methods for Avoiding Known Dataset Biases](https://mlsanity.com/paper/1909.03683/) (2019)
- [Counterfactual VQA: A Cause-Effect Look at Language Bias](https://mlsanity.com/paper/2006.04315/) (2020)
- [Image Retrieval on Real-life Images with Pre-trained Vision-and-Language Models](https://mlsanity.com/paper/2108.04024/) (2021)
- [Composing Text and Image for Image Retrieval - An Empirical Odyssey](https://mlsanity.com/paper/1812.07119/) (2018)
- [Qwen3-VL-Embedding and Qwen3-VL-Reranker: A Unified Framework for State-of-the-Art Multimodal Retrieval and Ranking](https://mlsanity.com/paper/2601.04720/) (2026)
- [RUBi: Reducing Unimodal Biases in Visual Question Answering](https://mlsanity.com/paper/1906.10169/) (2019)
- [Zero-Shot Composed Image Retrieval with Textual Inversion](https://mlsanity.com/paper/2303.15247/) (2023)
- [GME: Improving Universal Multimodal Retrieval by Multimodal LLMs](https://mlsanity.com/paper/2412.16855/) (2024)
- [UniIR: Training and Benchmarking Universal Multimodal Information Retrievers](https://mlsanity.com/paper/2311.17136/) (2023)
- [MM-Embed: Universal Multimodal Retrieval with Multimodal LLMs](https://mlsanity.com/paper/2411.02571/) (2024)
- [Pic2Word: Mapping Pictures to Words for Zero-shot Composed Image Retrieval](https://mlsanity.com/paper/2302.03084/) (2023)
- [MagicLens: Self-Supervised Image Retrieval with Open-Ended Instructions](https://mlsanity.com/paper/2403.19651/) (2024)
- [VLM2Vec-V2: Advancing Multimodal Embedding for Videos, Images, and Visual Documents](https://mlsanity.com/paper/2507.04590/) (2025)
- [LamRA: Large Multimodal Model as Your Advanced Retrieval Assistant](https://mlsanity.com/paper/2412.01720/) (2024)
- [Overcoming Language Priors in Visual Question Answering with Adversarial Regularization](https://mlsanity.com/paper/1810.03649/) (2018)
- [Vision-by-Language for Training-Free Compositional Image Retrieval](https://mlsanity.com/paper/2310.09291/) (2023)
- [Characterizing and overcoming the greedy nature of learning in multi-modal deep neural networks](https://mlsanity.com/paper/2202.05306/) (2022)
- [CompoDiff: Versatile Composed Image Retrieval With Latent Diffusion](https://mlsanity.com/paper/2303.11916/) (2023)
- [Vision-and-Language or Vision-for-Language? On Cross-Modal Influence in Multimodal Transformers](https://mlsanity.com/paper/2109.04448/) (2021)
- [Beyond Question-Based Biases: Assessing Multimodal Shortcut Learning in Visual Question Answering](https://mlsanity.com/paper/2104.03149/) (2021)
- [MIRAGE: The Illusion of Visual Understanding](https://mlsanity.com/paper/2603.21687/) (2026)
- [Does my multimodal model learn cross-modal interactions? It's harder to tell than you might think!](https://mlsanity.com/paper/2010.06572/) (2020)
- [Compositional Learning of Image-Text Query for Image Retrieval](https://mlsanity.com/paper/2006.11149/) (2020)

## Abstract (arXiv)

Composed Image Retrieval (CIR) is a multimodal retrieval task where a query consists of a reference image and a textual modification, and the goal is to retrieve a target image satisfying both. In principle, strong performance on CIR benchmarks is assumed to require multimodal composition, i.e., combining complementary information from reference image and textual modification. In this work, we show that this assumption does not always hold. Across four widely used CIR benchmarks and eleven Generalist Multimodal Embedding models, a large fraction of queries can be solved using a single modality (from 32.2% to 83.6%), revealing pervasive unimodal shortcuts. Thus, high CIR performance can arise from unimodal signals rather than true multimodal composition. To better understand this issue, we perform a two-stage audit. First, we identify shortcut-solvable queries through cross-model analysis. Second, we conduct human validation on 4,741 shortcut-free queries, of which only 1,689 are well-formed, with common issues including ambiguous edits and mismatched targets. Re-evaluating models on this validated subset reveals qualitatively different behaviour: queries can no longer be solved with a single modality, and successful retrieval requires combining both inputs. While accuracy decreases, reliance on multimodal information increases. Overall, current CIR benchmarks conflate shortcut-solvable, noisy, and genuinely compositional queries, leading to an overestimation of model capability in multimodal composition.