VisMin: Visual Minimal-Change Understanding

Fine-grained understanding of objects, attributes, and relationships between objects is crucial for visual-language models (VLMs). Existing benchmarks primarily focus on evaluating VLMs' capability to distinguish between two very similar \textit{captions} given an image. In this paper, we introduce a new, challenging benchmark termed \textbf{Vis}ual \textbf{Min}imal-Change Understanding (VisMin), which requires models to predict the correct image-caption match given two images and two captions. The image pair and caption pair contain minimal changes, i.e., only one aspect changes at a time from among the following: \textit{object}, \textit{attribute}, \textit{count}, and \textit{spatial relation}. These changes test the models' understanding of objects, attributes (such as color, material, shape), counts, and spatial relationships between objects. We built an automatic framework using large language models and diffusion models, followed by a rigorous 4-step verification process by human annotators. Empirical experiments reveal that current VLMs exhibit notable deficiencies in understanding spatial relationships and counting abilities. We also generate a large-scale training dataset to finetune CLIP and Idefics2, showing significant improvements in fine-grained understanding across benchmarks and in CLIP's general image-text alignment. We release all resources, including the benchmark, training data, and finetuned model checkpoints, at \url{https://vismin.net/}.

Benchmarking Vision Language Models for Cultural Understanding

Foundation models and vision-language pre-training have notably advanced Vision Language Models (VLMs), enabling multimodal processing of visual and linguistic data. However, their performance has been typically assessed on general scene understanding - recognizing objects, attributes, and actions - rather than cultural comprehension. This study introduces CulturalVQA, a visual question-answering benchmark aimed at assessing VLM's geo-diverse cultural understanding. We curate a collection of 2,378 image-question pairs with 1-5 answers per question representing cultures from 11 countries across 5 continents. The questions probe understanding of various facets of culture such as clothing, food, drinks, rituals, and traditions. Benchmarking VLMs on CulturalVQA, including GPT-4V and Gemini, reveals disparity in their level of cultural understanding across regions, with strong cultural understanding capabilities for North America while significantly lower performance for Africa. We observe disparity in their performance across cultural facets too, with clothing, rituals, and traditions seeing higher performances than food and drink. These disparities help us identify areas where VLMs lack cultural understanding and demonstrate the potential of CulturalVQA as a comprehensive evaluation set for gauging VLM progress in understanding diverse cultures.

Decompose and Compare Consistency: Measuring VLMs' Answer Reliability via Task-Decomposition Consistency Comparison

Despite tremendous advancements, current state-of-the-art Vision-Language Models (VLMs) are still far from perfect. They tend to hallucinate and may generate biased responses. In such circumstances, having a way to assess the reliability of a given response generated by a VLM is quite useful. Existing methods, such as estimating uncertainty using answer likelihoods or prompt-based confidence generation, often suffer from overconfidence. Other methods use self-consistency comparison but are affected by confirmation biases. To alleviate these, we propose \textbf{De}compose and \textbf{C}ompare \textbf{C}onsistency (\texttt{DeCC}) for reliability measurement. By comparing the consistency between the direct answer generated using the VLM's internal reasoning process, and the indirect answers obtained by decomposing the question into sub-questions and reasoning over the sub-answers produced by the VLM, \texttt{DeCC} measures the reliability of VLM's direct answer. Experiments across six vision-language tasks with three VLMs show \texttt{DeCC}'s reliability estimation achieves better correlation with task accuracy compared to the existing methods.

An Introduction to Vision-Language Modeling

Following the recent popularity of Large Language Models (LLMs), several attempts have been made to extend them to the visual domain. From having a visual assistant that could guide us through unfamiliar environments to generative models that produce images using only a high-level text description, the vision-language model (VLM) applications will significantly impact our relationship with technology. However, there are many challenges that need to be addressed to improve the reliability of those models. While language is discrete, vision evolves in a much higher dimensional space in which concepts cannot always be easily discretized. To better understand the mechanics behind mapping vision to language, we present this introduction to VLMs which we hope will help anyone who would like to enter the field. First, we introduce what VLMs are, how they work, and how to train them. Then, we present and discuss approaches to evaluate VLMs. Although this work primarily focuses on mapping images to language, we also discuss extending VLMs to videos.

Improving Text-to-Image Consistency via Automatic Prompt Optimization

Impressive advances in text-to-image (T2I) generative models have yielded a plethora of high performing models which are able to generate aesthetically appealing, photorealistic images. Despite the progress, these models still struggle to produce images that are consistent with the input prompt, oftentimes failing to capture object quantities, relations and attributes properly. Existing solutions to improve prompt-image consistency suffer from the following challenges: (1) they oftentimes require model fine-tuning, (2) they only focus on nearby prompt samples, and (3) they are affected by unfavorable trade-offs among image quality, representation diversity, and prompt-image consistency. In this paper, we address these challenges and introduce a T2I optimization-by-prompting framework, OPT2I, which leverages a large language model (LLM) to improve prompt-image consistency in T2I models. Our framework starts from a user prompt and iteratively generates revised prompts with the goal of maximizing a consistency score. Our extensive validation on two datasets, MSCOCO and PartiPrompts, shows that OPT2I can boost the initial consistency score by up to 24.9% in terms of DSG score while preserving the FID and increasing the recall between generated and real data. Our work paves the way toward building more reliable and robust T2I systems by harnessing the power of LLMs.

MoqaGPT : Zero-Shot Multi-modal Open-domain Question Answering with Large Language Model

Multi-modal open-domain question answering typically requires evidence retrieval from databases across diverse modalities, such as images, tables, passages, etc. Even Large Language Models (LLMs) like GPT-4 fall short in this task. To enable LLMs to tackle the task in a zero-shot manner, we introduce MoqaGPT, a straightforward and flexible framework. Using a divide-and-conquer strategy that bypasses intricate multi-modality ranking, our framework can accommodate new modalities and seamlessly transition to new models for the task. Built upon LLMs, MoqaGPT retrieves and extracts answers from each modality separately, then fuses this multi-modal information using LLMs to produce a final answer. Our methodology boosts performance on the MMCoQA dataset, improving F1 by +37.91 points and EM by +34.07 points over the supervised baseline. On the MultiModalQA dataset, MoqaGPT surpasses the zero-shot baseline, improving F1 by 9.5 points and EM by 10.1 points, and significantly closes the gap with supervised methods. Our codebase is available at https://github.com/lezhang7/MOQAGPT.

Improving Automatic VQA Evaluation Using Large Language Models

8 years after the visual question answering (VQA) task was proposed, accuracy remains the primary metric for automatic evaluation. VQA Accuracy has been effective so far in the IID evaluation setting. However, our community is undergoing a shift towards open-ended generative models and OOD evaluation. In this new paradigm, the existing VQA Accuracy metric is overly stringent and underestimates the performance of VQA systems. Thus, there is a need to develop more robust automatic VQA metrics that serve as a proxy for human judgment. In this work, we propose to leverage the in-context learning capabilities of instruction-tuned large language models (LLMs) to build a better VQA metric. We formulate VQA evaluation as an answer-rating task where the LLM is instructed to score the accuracy of a candidate answer given a set of reference answers. We demonstrate the proposed metric better correlates with human judgment compared to existing metrics across several VQA models and benchmarks. We hope wide adoption of our metric will contribute to better estimating the research progress on the VQA task.

Contrasting Intra-Modal and Ranking Cross-Modal Hard Negatives to Enhance Visio-Linguistic Fine-grained Understanding

Current Vision and Language Models (VLMs) demonstrate strong performance across various vision-language tasks, yet they struggle with fine-grained understanding. This issue stems from weak image-caption alignment in pretraining datasets and a simplified contrastive objective that fails to distinguish nuanced grounding elements such as relations, actions, and attributes. As a result, the models tend to learn bag-of-words representations. To mitigate these challenges, we introduce an intra-modal contrastive loss and a unique cross-modal rank loss with an adaptive threshold that serves as curriculum learning, utilizing our automatically generated hard negatives to augment the model's capacity. Our strategy, which does not necessitate additional annotations or parameters, can be incorporated into any VLM trained with an image-text contrastive loss. Upon application to CLIP, our method leads to significant improvements on four fine-grained benchmarks, and it also enhances the performance of X-VLM, which is the state-of-art moodel on fine-grained reasoning.

Investigating Prompting Techniques for Zero- and Few-Shot Visual Question Answering

Visual question answering (VQA) is a challenging task that requires the ability to comprehend and reason with visual information. While recent vision-language models have made strides, they continue to struggle with zero-shot VQA, particularly in handling complex compositional questions and adapting to new domains i.e. knowledge-based reasoning. This paper explores the use of various prompting strategies, focusing on the BLIP2 model, to enhance zero-shot VQA performance. We conduct a comprehensive investigation across several VQA datasets, examining the effectiveness of different question templates, the role of few-shot exemplars, the impact of chain-of-thought (CoT) reasoning, and the benefits of incorporating image captions as additional visual cues. Despite the varied outcomes, our findings demonstrate that carefully designed question templates and the integration of additional visual cues, like image captions, can contribute to improved VQA performance, especially when used in conjunction with few-shot examples. However, we also identify a limitation in the use of chain-of-thought rationalization, which negatively affects VQA accuracy. Our study thus provides critical insights into the potential of prompting for improving zero-shot VQA performance.

An Examination of the Robustness of Reference-Free Image Captioning Evaluation Metrics

Recently, reference-free metrics such as CLIPScore (Hessel et al., 2021) and UMIC (Lee et al., 2021) have been proposed for automatic evaluation of image captions, demonstrating a high correlation with human judgment. In this work, our focus lies in evaluating the robustness of these metrics in scenarios that require distinguishing between two captions with high lexical overlap but very different meanings. Our findings reveal that despite their high correlation with human judgment, both CLIPScore and UMIC struggle to identify fine-grained errors in captions. However, when comparing different types of fine-grained errors, both metrics exhibit limited sensitivity to implausibility of captions and strong sensitivity to lack of sufficient visual grounding. Probing further into the visual grounding aspect, we found that both CLIPScore and UMIC are impacted by the size of image-relevant objects mentioned in the caption, and that CLIPScore is also sensitive to the number of mentions of image-relevant objects in the caption. In terms of linguistic aspects of a caption, we found that both metrics lack the ability to comprehend negation, UMIC is sensitive to caption lengths, and CLIPScore is insensitive to the structure of the sentence. We hope our findings will serve as a valuable guide towards improving reference-free evaluation in image captioning.