PyTorchKR Papers

“Visual Question Answering”

총 2,807편
태그·기간 등 상세 필터

카테고리 탐색

arXiv:2307.09886

A reinforcement learning approach for VQA validation: an application to diabetic macular edema grading

Recent advances in machine learning models have greatly increased the performance of automated methods in medical image analysis. However, the internal functioning of such models is largely hidden, which hinders their integration in clinical practice. Explainability and trust are viewed as important aspects of modern methods, for the latter's widespread use in clinical communities. As such, validation of machine learning models represents an important aspect and yet, most methods are only validated in a limited way. In this work, we focus on providing a richer and more appropriate validation approach for highly powerful Visual Question Answering (VQA) algorithms. To better understand the performance of these methods, which answer arbitrary questions related to images, this work focuses on an automatic visual Turing test (VTT). That is, we propose an automatic adaptive questioning method, that aims to expose the reasoning behavior of a VQA algorithm. Specifically, we introduce a reinforcement learning (RL) agent that observes the history of previously asked questions, and uses it to select the next question to pose. We demonstrate our approach in the context of evaluating algorithms that automatically answer questions related to diabetic macular edema (DME) grading. The experiments show that such an agent has similar behavior to a clinician, whereby asking questions that are relevant to key clinical concepts.

arXiv:2408.13898공식 코드 · Papers with Code

Evaluating Attribute Comprehension in Large Vision-Language Models

Currently, large vision-language models have gained promising progress on many downstream tasks. However, they still suffer many challenges in fine-grained visual understanding tasks, such as object attribute comprehension. Besides, there have been growing efforts on the evaluations of large vision-language models, but lack of in-depth study of attribute comprehension and the visual language fine-tuning process. In this paper, we propose to evaluate the attribute comprehension ability of large vision-language models from two perspectives: attribute recognition and attribute hierarchy understanding. We evaluate three vision-language interactions, including visual question answering, image-text matching, and image-text cosine similarity. Furthermore, we explore the factors affecting attribute comprehension during fine-tuning. Through a series of quantitative and qualitative experiments, we introduce three main findings: (1) Large vision-language models possess good attribute recognition ability, but their hierarchical understanding ability is relatively limited. (2) Compared to ITC, ITM exhibits superior capability in capturing finer details, making it more suitable for attribute understanding tasks. (3) The attribute information in the captions used for fine-tuning plays a crucial role in attribute understanding. We hope this work can help guide future progress in fine-grained visual understanding of large vision-language models.

arXiv:2410.10441공식 코드 · Papers with Code

Free Video-LLM: Prompt-guided Visual Perception for Efficient Training-free Video LLMs

Vision-language large models have achieved remarkable success in various multi-modal tasks, yet applying them to video understanding remains challenging due to the inherent complexity and computational demands of video data. While training-based video-LLMs deliver high performance, they often require substantial resources for training and inference. Conversely, training-free approaches offer a more efficient alternative by adapting pre-trained image-LLMs models for video tasks without additional training, but they face inference efficiency bottlenecks due to the large number of visual tokens generated from video frames. In this work, we present a novel prompt-guided visual perception framework (abbreviated as Free Video-LLM) for efficient inference of training-free video LLMs. The proposed framework decouples spatial-temporal dimension and performs temporal frame sampling and spatial RoI cropping respectively based on task-specific prompts. Our method effectively reduces the number of visual tokens while maintaining high performance across multiple video question-answering benchmarks. Extensive experiments demonstrate that our approach achieves competitive results with significantly fewer tokens, offering an optimal trade-off between accuracy and computational efficiency compared to state-of-the-art video LLMs. The code will be available at https://github.com/contrastive/FreeVideoLLM.

arXiv:2412.08169공식 코드 · Papers with Code

Illusory VQA: Benchmarking and Enhancing Multimodal Models on Visual Illusions

In recent years, Visual Question Answering (VQA) has made significant strides, particularly with the advent of multimodal models that integrate vision and language understanding. However, existing VQA datasets often overlook the complexities introduced by image illusions, which pose unique challenges for both human perception and model interpretation. In this study, we introduce a novel task called Illusory VQA, along with four specialized datasets: IllusionMNIST, IllusionFashionMNIST, IllusionAnimals, and IllusionChar. These datasets are designed to evaluate the performance of state-of-the-art multimodal models in recognizing and interpreting visual illusions. We assess the zero-shot performance of various models, fine-tune selected models on our datasets, and propose a simple yet effective solution for illusion detection using Gaussian and blur low-pass filters. We show that this method increases the performance of models significantly and in the case of BLIP-2 on IllusionAnimals without any fine-tuning, it outperforms humans. Our findings highlight the disparity between human and model perception of illusions and demonstrate that fine-tuning and specific preprocessing techniques can significantly enhance model robustness. This work contributes to the development of more human-like visual understanding in multimodal models and suggests future directions for adapting filters using learnable parameters.

arXiv:2104.00332

UC2: Universal Cross-lingual Cross-modal Vision-and-Language Pre-training

Vision-and-language pre-training has achieved impressive success in learning multimodal representations between vision and language. To generalize this success to non-English languages, we introduce UC2, the first machine translation-augmented framework for cross-lingual cross-modal representation learning. To tackle the scarcity problem of multilingual captions for image datasets, we first augment existing English-only datasets with other languages via machine translation (MT). Then we extend the standard Masked Language Modeling and Image-Text Matching training objectives to multilingual setting, where alignment between different languages is captured through shared visual context (i.e, using image as pivot). To facilitate the learning of a joint embedding space of images and all languages of interest, we further propose two novel pre-training tasks, namely Masked Region-to-Token Modeling (MRTM) and Visual Translation Language Modeling (VTLM), leveraging MT-enhanced translated data. Evaluation on multilingual image-text retrieval and multilingual visual question answering benchmarks demonstrates that our proposed framework achieves new state-of-the-art on diverse non-English benchmarks while maintaining comparable performance to monolingual pre-trained models on English tasks.

arXiv:2402.14683공식 코드 · Papers with Code

Visual Hallucinations of Multi-modal Large Language Models

Visual hallucination (VH) means that a multi-modal LLM (MLLM) imagines incorrect details about an image in visual question answering. Existing studies find VH instances only in existing image datasets, which results in biased understanding of MLLMs' performance under VH due to limited diversity of such VH instances. In this work, we propose a tool called VHTest to generate a diverse set of VH instances. Specifically, VHTest finds some initial VH instances in existing image datasets (e.g., COCO), generates a text description for each VH mode, and uses a text-to-image generative model (e.g., DALL-E-3) to generate VH images based on the text descriptions. We collect a benchmark dataset with 1,200 VH instances in 8 VH modes using VHTest. We find that existing MLLMs such as GPT-4V, LLaVA-1.5, and MiniGPT-v2 hallucinate for a large fraction of the instances in our benchmark. Moreover, we find that fine-tuning an MLLM using our benchmark dataset reduces its likelihood to hallucinate without sacrificing its performance on other benchmarks. Our benchmarks are publicly available: https://github.com/wenhuang2000/VHTest.

arXiv:2502.16989

All-in-one: Understanding and Generation in Multimodal Reasoning with the MAIA Benchmark

We introduce MAIA (Multimodal AI Assessment), a native-Italian benchmark designed for fine-grained investigation of the reasoning abilities of visual language models on videos. MAIA differs from other available video benchmarks for its design, its reasoning categories, the metric it uses and the language and culture of the videos. It evaluates Vision Language Models (VLMs) on two aligned tasks: a visual statement verification task, and an open-ended visual question-answering task, both on the same set of video-related questions. It considers twelve reasoning categories that aim to disentangle language and vision relations by highlight when one of two alone encodes sufficient information to solve the tasks, when they are both needed and when the full richness of the short video is essential instead of just a part of it. Thanks to its carefully taught design, it evaluates VLMs' consistency and visually grounded natural language comprehension and generation simultaneously through an aggregated metric. Last but not least, the video collection has been carefully selected to reflect the Italian culture and the language data are produced by native-speakers.

arXiv:2609.04242v1eess.AS

Training-Free Speech-Centric Omni Understanding with Frozen VLMs

Audio-visual understanding remains challenging because models must jointly interpret spoken content, visual events, and their temporal relationships. Existing omni models typically introduce dedicated audio encoders and rely on expensive audio-video-text training, tightly coupling omni capability to specific VLM backbones and potentially weakening their existing visual and reasoning abilities. This raises three questions: whether native omni training is necessary for every new VLM, whether speech-centric omni capability can be added while preserving the original backbone, and where richer acoustic representations remain essential. We introduce Training-Free Omni (TFO), a plug-and-play framework that converts a frozen VLM into a speech-centric omni model without architectural modification, or multimodal re-alignment. TFO uses Whisper to extract confidence-filtered, timestamped transcripts and routes them through the VLM's existing language interface, while leaving its visual pathway unchanged. Across matched comparisons with native omni models on 56 benchmarks and 21 languages, TFO is competitive on audio-visual understanding, improves average audio-only performance across all five model settings, and achieves substantial multilingual speech gains. Freezing the VLM also generally preserves stronger image/video understanding, visual grounding, coding, mathematical reasoning, and medical question answering than the corresponding native omni checkpoints. These results show that strong speech-centric omni understanding can often be obtained through modular audio-to-language routing rather than costly backbone-specific training.

arXiv:2310.18235

Davidsonian Scene Graph: Improving Reliability in Fine-grained Evaluation for Text-to-Image Generation

Evaluating text-to-image models is notoriously difficult. A strong recent approach for assessing text-image faithfulness is based on QG/A (question generation and answering), which uses pre-trained foundational models to automatically generate a set of questions and answers from the prompt, and output images are scored based on whether these answers extracted with a visual question answering model are consistent with the prompt-based answers. This kind of evaluation is naturally dependent on the quality of the underlying QG and VQA models. We identify and address several reliability challenges in existing QG/A work: (a) QG questions should respect the prompt (avoiding hallucinations, duplications, and omissions) and (b) VQA answers should be consistent (not asserting that there is no motorcycle in an image while also claiming the motorcycle is blue). We address these issues with Davidsonian Scene Graph (DSG), an empirically grounded evaluation framework inspired by formal semantics, which is adaptable to any QG/A frameworks. DSG produces atomic and unique questions organized in dependency graphs, which (i) ensure appropriate semantic coverage and (ii) sidestep inconsistent answers. With extensive experimentation and human evaluation on a range of model configurations (LLM, VQA, and T2I), we empirically demonstrate that DSG addresses the challenges noted above. Finally, we present DSG-1k, an open-sourced evaluation benchmark that includes 1,060 prompts, covering a wide range of fine-grained semantic categories with a balanced distribution. We release the DSG-1k prompts and the corresponding DSG questions.

arXiv:2302.04865

ELBA: Learning by Asking for Embodied Visual Navigation and Task Completion

The research community has shown increasing interest in designing intelligent embodied agents that can assist humans in accomplishing tasks. Although there have been significant advancements in related vision-language benchmarks, most prior work has focused on building agents that follow instructions rather than endowing agents the ability to ask questions to actively resolve ambiguities arising naturally in embodied environments. To address this gap, we propose an Embodied Learning-By-Asking (ELBA) model that learns when and what questions to ask to dynamically acquire additional information for completing the task. We evaluate ELBA on the TEACh vision-dialog navigation and task completion dataset. Experimental results show that the proposed method achieves improved task performance compared to baseline models without question-answering capabilities.

arXiv:2406.03015

Balancing Performance and Efficiency in Zero-shot Robotic Navigation

We present an optimization study of the Vision-Language Frontier Maps (VLFM) applied to the Object Goal Navigation task in robotics. Our work evaluates the efficiency and performance of various vision-language models, object detectors, segmentation models, and multi-modal comprehension and Visual Question Answering modules. Using the $\textit{val-mini}$ and $\textit{val}$ splits of Habitat-Matterport 3D dataset, we conduct experiments on a desktop with limited VRAM. We propose a solution that achieves a higher success rate (+1.55%) improving over the VLFM BLIP-2 baseline without substantial success-weighted path length loss while requiring $\textbf{2.3 times}$ less video memory. Our findings provide insights into balancing model performance and computational efficiency, suggesting effective deployment strategies for resource-limited environments.

arXiv:2607.27084v4cs.CV

SciFigQual-Bench: 원고 연계 맥락을 활용한 과학 그림 품질 평가

SciFigQual-Bench: Scientific Figure Quality Assessment with Manuscript-Linked Context

과학 그림의 품질은 원고와 결부되어 있다. 잘라낸 그림이 시각적으로 선명하더라도 그 캡션이나 이를 인용하는 문단과 모순될 수 있다. 대부분의 이미지 품질 평가 및 차트 이해 방법은 그림, 캡션, 인용 텍스트 간의 정합성이 아니라, 분리된 시각적 표면이나 질의응답 점수를 대상으로 한다. 이미지와 텍스트를 함께 보는 단일 종단간 프롬프트는 여전히 눈에 보이는 것과 캡션 및 인용 문단이 주장하는 것을 뒤섞기 때문에, 유창한 문구가 시각 점수를 높일 수 있고 텍스트가 없으면 낮은 품질로 취급될 수 있다. 이러한 간극을 해소하기 위해 본 논문은 SciFigQual-Bench를 제안한다. 이는 출판된 CS 학회 논문의 그림을 대상으로 하는 원고 연계 벤치마크로, 각 그림을 캡션 및 색인으로 확인된 인용 문단과 연결하고, 시각적 명료성, 레이아웃, 캡션 일관성, 맥락 일관성, 오해 유발 위험을 점수화하며, 근거가 없는 차원은 평가하지 않은 채로 둔다. SFQ-Agent는 이미지와 텍스트를 별도의 호출로 읽고 모달리티별 근거를 기록한다. 교차 모달 판정기는 이 보고서들로부터 캡션 및 인용 텍스트와의 정합성을 점수화하고, 결정론적 실행기는 시각 점수를 유지하며 고정된 규칙으로 오해 유발 위험을 정하고 명시된 상한을 적용하므로, 각 차원은 단일 종단간 프롬프트가 아니라 평가 기준을 따른다. 이 단계별 판정기를 단일 패스 및 OCR 보조 프로토콜과 비교한 실험에서, 단계별 판정 하에서 인간 평가 평균에 대한 점추정 적합도가 가장 가까웠으나 프로토콜 간 차이는 작게 유지되었다. 캡션 일관성은 여전히 주요한 격차로 남아 있으며, 평가자 평균과의 일치도는 인간 평가자 간 일치도와는 다른 질문에 답한다. 프로젝트 페이지: https://frankdengai.github.io/SciFigQual-Bench/

arXiv:2111.10023

UFO: A UniFied TransfOrmer for Vision-Language Representation Learning

In this paper, we propose a single UniFied transfOrmer (UFO), which is capable of processing either unimodal inputs (e.g., image or language) or multimodal inputs (e.g., the concatenation of the image and the question), for vision-language (VL) representation learning. Existing approaches typically design an individual network for each modality and/or a specific fusion network for multimodal tasks. To simplify the network architecture, we use a single transformer network and enforce multi-task learning during VL pre-training, which includes the image-text contrastive loss, image-text matching loss, and masked language modeling loss based on the bidirectional and the seq2seq attention mask. The same transformer network is used as the image encoder, the text encoder, or the fusion network in different pre-training tasks. Empirically, we observe less conflict among different tasks and achieve new state of the arts on visual question answering, COCO image captioning (cross-entropy optimization) and nocaps (in SPICE). On other downstream tasks, e.g., image-text retrieval, we also achieve competitive performance.

arXiv:2405.08813

CinePile: A Long Video Question Answering Dataset and Benchmark

Current datasets for long-form video understanding often fall short of providing genuine long-form comprehension challenges, as many tasks derived from these datasets can be successfully tackled by analyzing just one or a few random frames from a video. To address this issue, we present a novel dataset and benchmark, CinePile, specifically designed for authentic long-form video understanding. This paper details our innovative approach for creating a question-answer dataset, utilizing advanced LLMs with human-in-the-loop and building upon human-generated raw data. Our comprehensive dataset comprises 305,000 multiple-choice questions (MCQs), covering various visual and multimodal aspects, including temporal comprehension, understanding human-object interactions, and reasoning about events or actions within a scene. Additionally, we fine-tuned open-source Video-LLMs on the training split and evaluated both open-source and proprietary video-centric LLMs on the test split of our dataset. The findings indicate that although current models underperform compared to humans, fine-tuning these models can lead to significant improvements in their performance.

arXiv:2610.17937v1cs.CV

MERIT: Multimodal Evidence Reasoning over Image Tables

Table reasoning with language models typically assumes access to clean textual tables. In practice, tables are often serialized into text, which can corrupt cell contents, distort structural relationships, and discard visual cues, leaving the model to reason over incomplete evidence. In contrast, table images retain this information, but using it requires both accurate perception and reasoning grounded in the relevant cells. To address this, we propose MERIT (Multimodal Evidence Reasoning over Image Tables), a three-stage framework for learning to reason over table images, with serialized tables serving as optional auxiliary inputs (if provided). Our analysis identifies visual evidence acquisition as a bottleneck, motivating table-specific encoder adaptation through reconstruction, localization, and color-conditioned supervision. Building on the adapted representations, supervised fine-tuning teaches explicit evidence selection and reasoning using verified trajectories. Finally, reinforcement learning refines the policy with separate rewards for evidence correctness, calculation, final answers, and output format. By providing feedback on intermediate steps, the framework targets retrieval and computational errors that final-answer supervision alone does not distinguish. We evaluate the approach on table question answering and fact verification, analyze the contributions of visual inputs and training components, and assess generalization to additional datasets with different table structures and visual presentations.

arXiv:2404.12803

TextSquare: Scaling up Text-Centric Visual Instruction Tuning

Text-centric visual question answering (VQA) has made great strides with the development of Multimodal Large Language Models (MLLMs), yet open-source models still fall short of leading models like GPT4V and Gemini, partly due to a lack of extensive, high-quality instruction tuning data. To this end, we introduce a new approach for creating a massive, high-quality instruction-tuning dataset, Square-10M, which is generated using closed-source MLLMs. The data construction process, termed Square, consists of four steps: Self-Questioning, Answering, Reasoning, and Evaluation. Our experiments with Square-10M led to three key findings: 1) Our model, TextSquare, considerably surpasses open-source previous state-of-the-art Text-centric MLLMs and sets a new standard on OCRBench(62.2%). It even outperforms top-tier models like GPT4V and Gemini in 6 of 10 text-centric benchmarks. 2) Additionally, we demonstrate the critical role of VQA reasoning data in offering comprehensive contextual insights for specific questions. This not only improves accuracy but also significantly mitigates hallucinations. Specifically, TextSquare scores an average of 75.1% across four general VQA and hallucination evaluation datasets, outperforming previous state-of-the-art models. 3) Notably, the phenomenon observed in scaling text-centric VQA datasets reveals a vivid pattern: the exponential increase of instruction tuning data volume is directly proportional to the improvement in model performance, thereby validating the necessity of the dataset scale and the high quality of Square-10M.

arXiv:2004.12070공식 코드 · Papers with Code

Deep Multimodal Neural Architecture Search

Designing effective neural networks is fundamentally important in deep multimodal learning. Most existing works focus on a single task and design neural architectures manually, which are highly task-specific and hard to generalize to different tasks. In this paper, we devise a generalized deep multimodal neural architecture search (MMnas) framework for various multimodal learning tasks. Given multimodal input, we first define a set of primitive operations, and then construct a deep encoder-decoder based unified backbone, where each encoder or decoder block corresponds to an operation searched from a predefined operation pool. On top of the unified backbone, we attach task-specific heads to tackle different multimodal learning tasks. By using a gradient-based NAS algorithm, the optimal architectures for different tasks are learned efficiently. Extensive ablation studies, comprehensive analysis, and comparative experimental results show that the obtained MMnasNet significantly outperforms existing state-of-the-art approaches across three multimodal learning tasks (over five datasets), including visual question answering, image-text matching, and visual grounding.

arXiv:2505.24227

Light as Deception: GPT-driven Natural Relighting Against Vision-Language Pre-training Models

While adversarial attacks on vision-and-language pretraining (VLP) models have been explored, generating natural adversarial samples crafted through realistic and semantically meaningful perturbations remains an open challenge. Existing methods, primarily designed for classification tasks, struggle when adapted to VLP models due to their restricted optimization spaces, leading to ineffective attacks or unnatural artifacts. To address this, we propose \textbf{LightD}, a novel framework that generates natural adversarial samples for VLP models via semantically guided relighting. Specifically, LightD leverages ChatGPT to propose context-aware initial lighting parameters and integrates a pretrained relighting model (IC-light) to enable diverse lighting adjustments. LightD expands the optimization space while ensuring perturbations align with scene semantics. Additionally, gradient-based optimization is applied to the reference lighting image to further enhance attack effectiveness while maintaining visual naturalness. The effectiveness and superiority of the proposed LightD have been demonstrated across various VLP models in tasks such as image captioning and visual question answering.

arXiv:2004.06698공식 코드 · Papers with Code

Reasoning Visual Dialog with Sparse Graph Learning and Knowledge Transfer

Visual dialog is a task of answering a sequence of questions grounded in an image using the previous dialog history as context. In this paper, we study how to address two fundamental challenges for this task: (1) reasoning over underlying semantic structures among dialog rounds and (2) identifying several appropriate answers to the given question. To address these challenges, we propose a Sparse Graph Learning (SGL) method to formulate visual dialog as a graph structure learning task. SGL infers inherently sparse dialog structures by incorporating binary and score edges and leveraging a new structural loss function. Next, we introduce a Knowledge Transfer (KT) method that extracts the answer predictions from the teacher model and uses them as pseudo labels. We propose KT to remedy the shortcomings of single ground-truth labels, which severely limit the ability of a model to obtain multiple reasonable answers. As a result, our proposed model significantly improves reasoning capability compared to baseline methods and outperforms the state-of-the-art approaches on the VisDial v1.0 dataset. The source code is available at https://github.com/gicheonkang/SGLKT-VisDial.

arXiv:2610.00953v1cs.CV

디퓨전 MLLM의 두 가지 시계: 추론 근거가 전개되기 전에 답변이 안정화되는 시점

Two Clocks in Diffusion MLLMs: When Answers Stabilize Before Rationales Unfold

마스크드 디퓨전(masked diffusion) MLLM에서 답변 후보는 그 추론 근거(rationale)가 여전히 전개되는 동안에도 안정화될 수 있습니다. 본 연구에서는 기록된 후보의 사후적(retrospective) 안정화와 토큰 확정(token commitment)을 구분하고, 근거 생성과 관련하여 이 두 가지 시계를 조사합니다. 세 가지 시각적 질의응답 벤치마크 전반에서 결과를 분석한 결과, 단일 블록 및 EOS 억제 설정의 LaViDa 실행 시 안정화 시점에서 근거 측 캔버스의 89.4~98.1%가 작성되지 않은 채로 남아 있음을 발견했습니다. V*Bench에서는 블록 길이를 128에서 8로 줄임에 따라 답변 커버리지 및 관찰 가능 윈도우와 함께 이 비율이 89.4%에서 1.7%로 변화합니다. EOS 활성화 프롬프팅 환경에서 직접적인 지시사항은 M3CoT 및 ScienceQA에서 Nemotron의 전체 정확도를 각각 15.0 및 19.5%포인트 향상시키지만, LaViDa/V*Bench 정확도는 11.0%포인트 감소시킵니다. 대칭적 분해 분석을 통해 각 변화에서 더 큰 절대적 구성 요소가 조건부 정확도보다는 커버리지와 관련됨을 보였습니다. 캔버스를 일치시킨 이미지 어블레이션(ablation)은 답변 안정화와 함께 시각적 민감도를 측정하여 두 시간적 판독값을 분리합니다. 종합적으로, 이러한 측정들은 답변 안정화, 근거 전개, 시각적 민감도를 명확히 구분하며, 프롬프팅 차이에서 더 큰 비중을 차지하는 요소가 커버리지임을 규명합니다.