论文检索

Xiangjian Zeng, Wenjing Li, Qingqiang Wu, Liang Zhang

This paper presents FDC-Ground, a reinforcement learning framework that addresses the high-cost, low-signal challenge of GUI grounding training. The framework introduces two core contributions: (1) the Exponentially Decayed Distance Reward (EDDR), which provides resolution-robust and continuous feedback for position predictions, and (2) the Fact-Aligned Dynamic Completions Pruning (FDC-Pruning) strategy, which selectively retains completions whose advantage signs align with factual correctness, thereby reducing computational overhead while enhancing gradient quality and training stability. Using only 3.2K training samples and a single epoch, our 7B-parameter model achieves 88.3% and 91.0% accuracy on ScreenSpot and ScreenSpot-v2, outperforming several RL-based models such as UIShift and SE-GUI. Our 3B-parameter model based on Qwen2.5-VL-3B surpasses its original performance by +26.6%, demonstrating the effectiveness of our reward design and pruning strategy under low-resource conditions. Furthermore, the proposed FDC-Pruning strategy achieves a 1.18× training speedup and a +5.9% accuracy improvement over standard GRPO, and expanding the exploration space to 4× yields an additional +10.5% gain, confirming both the scalability and the training efficiency of our approach. These findings highlight that combining EDDR with FDC-Pruning offers a practical path toward scalable and efficient RL-based GUI grounding, even in low-resource settings.

Li Yuan, Qingfei Huang, Bingshan Zhu, Yi Cai, Qingbao Huang, Changmeng Zheng, Zikun Deng, Tao Wang

Multimodal Knowledge Editing (MKE) extends traditional knowledge editing to settings involving both textual and visual modalities. However, existing MKE benchmarks primarily assess final answer correctness, neglecting the quality of intermediate reasoning and robustness to visually rephrased inputs. To address this limitation, we introduce MMQAKE, the first benchmark for multimodal multihop question answering with knowledge editing. MMQAKE evaluates: (1) a model’s ability to reason over 2–5-hop factual chains that span both text and images, including performance at each intermediate step; (2) robustness to visually rephrased inputs in multihop questions. Our evaluation shows that current MKE methods often struggle to consistently update and reason over multimodal reasoning chains following knowledge edits. To overcome these challenges, we propose Hybrid-DMKG, a hybrid reasoning framework built on a dynamic multimodal knowledge graph (DMKG) to enable accurate multihop reasoning over updated multimodal knowledge. Hybrid-DMKG first uses a large language model to decompose multimodal multihop questions into sequential sub-questions, then applies a multimodal retrieval model to locate updated facts by jointly encoding each sub-question with candidate entities and their associated images. For answer inference, a hybrid reasoning module operates over the DMKG via two parallel paths: (1) relation-linking prediction; (2) RAG Reasoning with large vision-language models. A background-reflective decision module then aggregates evidence from both paths to select the most credible answer. Experimental results on MMQAKE show that Hybrid-DMKG significantly outperforms existing MKE approaches, achieving higher accuracy and improved robustness to knowledge updates.

Shangshang Yang, Xuewen Duan, Xiaoshan Yu, Ziwen Wang, Haiping Ma, Xingyi Zhang

Cognitive diagnosis (CD), inferring student knowledge mastery based on historical response records, is crucial for personalized educational services such as adaptive practice and learning path planning. Existing CD models were built based on the assumption that student's response data is integral, overlooking the nonrandom missingness of data caused by student answering exercises selectively. This missingness generally leads to biased and incomplete observations, where confounders, such as selection bias and exposure bias, significantly undermine the accuracy of student knowledge modeling. To address missingness, we propose a Debiased Cognitive Diagnosis (DBCD) framework through the perspective of counterfactual modeling to remove exogenous confounders from the response data. Specifically, the proposed DBCD achieves debiasing for CD by applying the idea of contrastive learning to constrain the model's prediction distributions on both factual and counterfactual data. For a student, the factual data is his/her original response records, while the counterfactual data is generated by sampling the same number of exercises from all exercises of each concept through a similarity-based counterfactual sampling strategy. Considering the difficulty of directly removing the exogenous confounders for student, we devise a β-Variational Autoencoder to model their exogenous confounders within the latent representations of knowledge proficiency by leveraging exercise priors and student response patterns. Then, the learned representations are further combined with the vanilla student's ability embedding via a gating mechanism-based fusion for final diagnosis prediction of the model. Extensive experiments on real-world educational datasets demonstrate that the proposed DBCD effectively mitigates confounders and even outperforms existing methods, thereby validating the feasibility and effectiveness of the DBCD framework.

Xiangyu Wang, Yanze Gao, Changxin Rong, Lyuzhou Chen, Derui Lyu, Xiren Zhou, Taiyu Ban, Huanhuan Chen

Zero-shot classifier expansion aims to adapt existing model to new, unseen classes. It utilizes class attributes or textual descriptions to learn a mapping from the semantic space to the classifier's weight space, without requiring new visual training data. However, the learning process for this mapping relies solely on correlating semantic patterns with their corresponding classifier weights and lacks explicit modeling of inter-class differences. This makes it difficult for the model to capture the critical discriminative features required to define classification boundaries. To overcome this limitation, we reframe the problem from a causal perspective and introduce a novel framework driven by counterfactuals. Our method first generates factual descriptions alongside corresponding inter-class counterfactuals to pinpoint the causal attributes essential for classification, then refines these representations via a mutual purification process, and finally leverages a novel separation loss to explicitly push the factual and counterfactual classifier weights apart. This strategy forces the model to forge clearer and more discriminative classification boundaries, achieving more accurate and robust classification. Extensive experiments demonstrate that our approach significantly outperforms existing state-of-the-art methods.

Masahiro Mitani, Yuya Sasaki

Temporal Graph Neural Network (TGNN) explanation has attracted increasing attention due to its applicability in dynamic scenarios such as recommendation systems. However, existing explanation methods for TGNNs face two key limitations: (1) computational inefficiency and (2) a restricted focus on either factual or counterfactual explanations, but not both. In this paper, we propose QIEA-TGX, an efficient and unified explanation algorithm based on a quantum-inspired evolutionary algorithm. QIEA-TGX effectively generates explanatory subgraphs that significantly influence TGNN predictions, without requiring additional model training or extensive inference. Experimental results on real-world datasets demonstrate that QIEA-TGX improves explanation fidelity by up to 31% while reducing computation time by up to 92% compared to state-of-the-art baselines.

Yifan Jia, Yuntao Du, Kailin Jiang, Yuyang Liang, Qihan Ren, Yi Xin, Rui Yang, Fenze Feng, MingCai Chen, Hengyang Lu 等

Large Multimodal Models (LMMs) face notable challenges when encountering multimodal knowledge conflicts, particularly under retrieval-augmented generation (RAG) frameworks, where the contextual information from external sources may contradict the model’s internal parametric knowledge, leading to unreliable outputs. However, existing benchmarks fail to reflect such realistic conflict scenarios. Most focus solely on intra-memory conflicts, while context-memory and inter-context conflicts remain largely unaddressed. Furthermore, commonly used factual knowledge-based evaluations are often overlooked, and existing datasets lack a thorough investigation into conflict detection capabilities.To bridge this gap, we propose MMKC-Bench, a benchmark designed to evaluate factual knowledge conflicts in both context-memory and inter-context scenarios. MMKC-Bench encompasses four types of multimodal knowledge conflicts and includes 1,881 knowledge instances and 3,997 images across 32 broad types, collected through automated pipelines with human verification. We evaluate four representative series of LMMs on both model behavior analysis and conflict detection tasks. Our findings show that while current LMMs are capable of recognizing knowledge conflicts, they tend to favor internal parametric knowledge over external evidence. We hope MMKC-Bench will foster further research in multimodal knowledge conflict and enhance the development of multimodal RAG systems.

Jinpeng Li, Hang Yu, Ziqi Ma, Peng Qi

Knowledge graph construction (KGC) aims to extract valuable information from text and organize it into structured knowledge graphs (KGs). Recent methods have leveraged the strong generative capabilities of large language models (LLMs) to improve the generalization and reduce the labor costs. However, constrained by the input length of LLMs, existing methods mainly focus on extracting knowledge within individual texts and lack the capability to discover latent knowledge across texts. To fill this gap, we propose a novel method for open knowledge graph construction, termed KG-DLF. The core idea of this method is to enhance the knowledge graph construction process by discovering new facts that are consistent with the underlying contextual logic. Specifically, we first design a knowledge extractor to extract knowledge from the text. Then, a knowledge normalizer performs schema alignment on the extracted knowledge. Next, we explore a knowledge discoverer based on a clue search strategy, which leverages the logical consistency of context to mine latent facts. Finally, we design a counterfactual-based knowledge corrector, enabling the model to purify knowledge and reduce factual errors. Experimental results show that KG-DLF is capable of extracting comprehensive knowledge in open-world scenarios across three KGC benchmarks.

Zhiqiang Liu, Yin Hua, Mingyang Chen, Yichi Zhang, Zhuo Chen, Lei Liang, Wen Zhang

Real-world knowledge graphs (KGs) contain not only standard triple-based facts, but also more complex, heterogeneous types of facts, such as hyper-relational facts with auxiliary key-value pairs, temporal facts with additional timestamps, and nested facts that imply relationships between facts. These richer forms of representation have attracted significant attention due to their enhanced expressiveness and capacity to model complex semantics in real-world scenarios. However, most existing studies suffer from two main limitations: (1) they typically focus on modeling only specific types of facts, thus making it difficult to generalize to real-world scenarios with multiple fact types; and (2) they struggle to achieve generalizable hierarchical (inter-fact and intra-fact) modeling due to the complexity of these representations. To overcome these limitations, we propose UniHR, a Unified Hierarchical Representation learning framework, which consists of a learning-optimized Hierarchical Data Representation (HiDR) module and a unified Hierarchical Structure Learning (HiSL) module. The HiDR module unifies hyper-relational KGs, temporal KGs, and nested factual KGs into triple-based representations. Then HiSL incorporates intra-fact and inter-fact message passing, focusing on enhancing both semantic information within individual facts and enriching the structural information between facts. To go beyond the unified method itself, we further explore the potential of unified representation in complex real-world scenarios. Extensive experiments on 9 datasets across 5 types of KGs demonstrate the effectiveness of UniHR and highlight the strong potential of unified representations.

Jilong Liu, Pengyang Shao, Wei Qin, Fei Liu, Yonghui Yang, Richang Hong

Knowledge Graph Question Answering (KGQA) aims to improve factual accuracy by leveraging structured knowledge. However, real-world Knowledge Graphs (KGs) are often incomplete, leading to the problem of Incomplete KGQA (IKGQA). A common solution is to incorporate external data to fill knowledge gaps, but existing methods lack the capacity to adaptively and contextually fuse multiple sources, failing to fully exploit their complementary strengths. To this end, we propose Debate over Mixed-knowledge (DoM), a novel framework that enables dynamic integration of structured and unstructured knowledge for IKGQA. Built upon the Multi-Agent Debate paradigm, DoM assigns specialized agents to perform inference over knowledge graphs and external texts separately, and coordinates their outputs through iterative interaction. It decomposes the input question into sub-questions, retrieves evidence via dual agents (KG and Retrieval-Augmented Generation, RAG), and employs a judge agent to evaluate and aggregate intermediate answers. This collaboration exploits knowledge complementarity and enhances robustness to KG incompleteness. In addition, existing IKGQA datasets simulate incompleteness by randomly removing triples, failing to capture the irregular and unpredictable nature of real-world knowledge incompleteness. To address this, we introduce a new dataset, Incomplete Knowledge Graph WebQuestions, constructed by leveraging real-world knowledge updates. These updates reflect knowledge beyond the static scope of KGs, yielding a more realistic and challenging benchmark. Through extensive experiments, we show that DoM consistently outperforms state-of-the-art baselines.

Chenxiao Lin, Ye Luo, KunHong Liu, Qingqiang Wu

Retrieval-augmented generation (RAG) enhances the reasoning capabilities of large language models (LLMs) by incorporating external knowledge. Among available sources, knowledge graphs (KGs) offer a structured and reliable foundation for factual information, making them increasingly popular in efforts to improve reasoning faithfulness in RAG. Most existing KG-based RAG methods rely on LLMs to extract knowledge from KGs. However, these approaches often require costly fine-tuning and struggle to navigate deep graph structures, limiting their effectiveness in multi-hop reasoning tasks. To address these challenges, we propose Stepwise Contrastive Reasoning (SCR), a lightweight framework that integrates graph structure and textual context for efficient and interpretable RAG over KGs. SCR combines relational message passing layers to encode KG entities with a Transformer encoder for processing question text. It decomposes reasoning into a series of alignment steps. At each step, SCR compares the current topic entity and its neighbors with the question representation, selecting the most relevant entity as the next topic entity. The question is then updated with this entity's textual description. This process continues until the selected entity no longer changes, indicating that the answer entity has been reached. Through stepwise alignment, SCR enables compact models to perform faithful and interpretable reasoning over large-scale KGs. Extensive experiments on several widely used KGQA benchmarks demonstrate that SCR not only achieves state-of-the-art performance but also effectively boosts the capabilities of smaller language models to match those of LLMs.

Lingyu Zhou, Yue Yu, Zhang Yi, Xiuyuan Xu

Entity hallucination poses a major challenge in radiology report generation (RRG), particularly for 3D CT scans where complex spatial contexts amplify factual errors. To address this, medical entity phrases serve as key carriers for multi-modal prompting, integrating expert knowledge into the vision-language model. Current methods use unified cross-attention for volume-phrase alignment, failing to account for anatomical specificity during the alignment process. In this work, we introduce the Dual-stream Entity Alignment Reporting network (DEAR) that separately models organ and lesion entities to resolve anatomical bias. Specifically, the dual-stream entity aligner is designed to partition medical entity phrases into organ and lesion streams, feeding them into separate cross-attention blocks in parallel to achieve fine-grained volume–phrase alignment. For structurally regular and spatially stable organ entities, an organ-guided cross-attention (OGCA) block is proposed to enforce structural consistency by retrieving the top-k voxel tokens via volume–phrase similarity and preserving spatial connectivity through morphological dilation. Meanwhile, a lesion-guided cross-attention (LGCA) block is introduced for structurally irregular and spatially variable lesion entities, enhancing anomaly sensitivity through phrase-weighted attention and refining discriminative boundaries via 3D residual Laplacian filtering. Experiments demonstrate that DEAR significantly reduces entity hallucinations and improves clinical factuality in 3D RRG benchmarks.

Zhaoyang Wei, Guoliang Wang, Guohua Gao, Yanchao Hao, Mingda Li, Wenchao Ding, Xi Chen, Shizhu He, Xuehui Yu

Although Vision Language Models (VLMs) have excelled at image and video understanding, applying them to hour-long videos is held back by two interrelated challenges: exorbitant computational expense and a qualitative breakdown in long-term temporal reasoning. Thus, models tend to generate answers based on speculation instead of solid visual facts, causing both factually incorrect and plausible hallucinations. This problem is compounded by current benchmarks that, by only emphasizing final answers, lack an effective mechanism to check whether reasoning is substantiated by specific visual evidence. This makes it hard to differentiate between true understanding and pretend comprehension, inhibiting targeted model refinement. To address these interrelated challenges of model fragility and evaluation weakness, we adopt a twofold strategy. First, we present EV²-Bench, a large-scale benchmark that breaks new ground by an evaluation paradigm built upon spatio-temporal visual evidence, forcing models to justify answers with checkable hints. Second, we put forward DynamicSelect, an adaptive token compression system that efficiently condenses salient information by a dynamic semantic selector and a hierarchical compression strategy. Comprehensive experiments demonstrate that DynamicSelect significantly outperforms the baselines on EV²-Bench as well as other public benchmarks. Our study offers not only a more effective approach to long-video understanding but also a more stringent evaluation paradigm, indicating the way toward more robust models.

Joao Alexandre Cardeira Pereira, Vasco Lopes, João C. Neves, David Semedo

Video Anomaly Understanding (VAU) is a novel task focused on describing unusual occurrences in videos. Despite growing interest, the evaluation of VAU remains an open challenge. Existing benchmarks rely on n-gram-based metrics (e.g., BLEU, ROUGE-L) or LLM-based evaluation. The first fails to capture the rich, free-form, and visually grounded nature of LVLM responses, while the latter focuses on assessing language quality over factual relevance, often resulting in subjective judgments that are misaligned with human perception. In this work, we address this issue by proposing FineVAU, a new benchmark for VAU that shifts the focus towards rich, fine-grained and domain-specific understanding of anomalous videos. We formulate VAU as a three-fold problem, with the goal of comprehensively understanding key descriptive elements of anomalies in video: events (What), participating entities (Who) and location (Where). Our benchmark introduces a) FV-Score, a novel, human-aligned evaluation metric that assesses the presence of critical visual elements in LVLM answers, providing interpretable, fine-grained feedback; and b) FineW³, a novel, comprehensive dataset curated through a structured and fully automatic procedure that augments existing human annotations with high quality, fine-grained visual information. Human evaluation reveals that our proposed metric has a superior alignment with human perception of anomalies in comparison to current approaches. Detailed experiments on FineVAU unveil critical limitations in LVLM's ability to perceive anomalous events that require spatial and fine-grained temporal understanding, despite strong performance on coarse grain, static information, and events with strong visual cues.

Xin Mei, Rui Mao, Xiaoyan Cai, Libin Yang, Erik Cambria

Automated interpretation and reporting of chest X-rays (CXRs) hold significant promise in reducing diagnostic errors and supporting radiologists under heavy clinical workloads. However, existing methods typically rely on global visual features and token-level supervision, limiting their sensitivity to subtle abnormalities and reducing their clinical reliability. To address these challenges, we present Reflective X-ray Network (RefleXNet), which systematically integrates multi-scale visual feature fusion and anatomical relational reasoning with a targeted self-reflective learning strategy. RefleXNet first constructs multi-scale visual representations and captures anatomical context through graph-based relational modeling. Building upon these representations, we introduce a targeted self-reflection strategy that uses clinically guided feedback from generated reports to selectively refine abnormality predictions and their associated region-level visual features. Extensive experiments on MIMIC-CXR demonstrate that RefleXNet consistently outperforms state-of-the-art baselines across clinical factual correctness metrics. Notably, our compact 3B-parameter model surpasses several recent models with over twice the parameter count. Additionally, RefleXNet exhibits strong generalization performance in zero-shot evaluations on IU-Xray compared with leading multimodal language models, highlighting its robustness and clinical effectiveness.

Xiao Liang, Chenxi Liu, Zhi Ma, Di Wang, Bin Jing, Quan Wang, Yuanyuan Shi

Medical Vision-Language Models (MedVLMs) show immense promise in clinical applicability. However, their reliability is hindered by hallucinations, where models often fail to derive answers from visual evidence, instead relying on learned textual priors. Existing mitigation strategies for MedVLMs have distinct limitations: training-based methods rely on costly expert annotations, limiting scalability, while training-free interventions like contrastive decoding, though data-efficient, apply a global, untargeted correction whose effects in complex real-world clinical settings can be unreliable. To address these challenges, we introduce Anatomical Region-Guided Contrastive Decoding (ARCD), a plug-and-play strategy that mitigates hallucinations by providing targeted, region-specific guidance. Our module leverages an anatomical mask to direct a three-tiered contrastive decoding process. By dynamically re-weighting at the token, attention, and logits levels, it verifiably steers the model's focus onto specified regions, reinforcing anatomical understanding and suppressing factually incorrect outputs. Extensive experiments across diverse datasets, including chest X-ray, CT, brain MRI, and ocular ultrasound, demonstrate our method's effectiveness in improving regional understanding, reducing hallucinations, and enhancing overall diagnostic accuracy.

Dayong Liang, Xiao-Yong Wei, Changmeng Zheng

Hallucination continues to pose a major obstacle in the reasoning capabilities of large language models (LLMs). Although the Multi-Agent Debate (MAD) paradigm offers a promising solution by promoting consensus among multiple agents to enhance reliability, it relies on the unrealistic assumption that all debaters are rational and reflective, which is a condition that may not hold when agents themselves are prone to hallucinations. To address this gap, we introduce the Multi-agent Undercover Gaming (MUG) protocol, inspired by social deduction games like ''Who is Undercover?''. MUG reframes MAD as a process of detecting ''undercover'' agents (those suffering from hallucinations) by employing multimodal counterfactual tests. Specifically, we modify reference images to introduce counterfactual evidence and observe whether agents can accurately identify these changes, providing ground-truth for identifying hallucinating agents and enabling robust, crowd-powered multimodal reasoning. MUG advances MAD protocols along three key dimensions: (1) enabling factual verification beyond statistical consensus through counterfactual testing; (2) introducing cross-evidence reasoning via dynamically modified evidence sources instead of relying on static inputs; and (3) fostering active reasoning, where agents engage in probing discussions rather than passively answering questions. Collectively, these innovations offer a more reliable and effective framework for multimodal reasoning in LLMs.

Meng Cao, Pengfei Hu, Yingyao Wang, Jihao Gu, Haoran Tang, Haoze Zhao, Chen Wang, Jiahua Dong, Wangbo Yu, Ge Zhang 等

Recent advancements in Large Video Language Models (LVLMs) have highlighted their potential for multi-modal understanding, yet evaluating their factual grounding in videos remains a critical unsolved challenge. To address this gap, we introduce Video SimpleQA, the first comprehensive benchmark tailored for factuality evaluation in video contexts. Our work differs from existing video benchmarks through the following key features: 1) Knowledge required: demanding integration of external knowledge beyond the video’s explicit narrative; 2) Multi-hop fact-seeking question: Each question involves multiple explicit facts and requires strict factual grounding without hypothetical or subjective inferences. We include per-hop single-fact-based sub-QAs alongside final QAs to enable fine-grained, step-by-step evaluation; 3) Short-form definitive answer: Answers are crafted as unambiguous and definitively correct in a short format with minimal scoring variance; 4) Temporal grounded required: Requiring answers to rely on one or more temporal segments in videos, rather than single frames. We extensively evaluate 33 state-of-the-art LVLMs and summarize key findings as follows: 1) Current LVLMs exhibit notable deficiencies in factual adherence, with the best-performing model o3 merely achieving an F-score of 66.3%; 2) Most LVLMs are overconfident in what they generate, with self-stated confidence exceeding actual accuracy; 3) Retrieval-Augmented Generation demonstrates consistent improvements at the cost of additional inference time overhead; 4) Multi-hop QA demonstrates substantially degraded performance compared to single-hop sub-QAs, with first-hop object/event recognition emerging as the primary bottleneck. We position Video SimpleQA as the cornerstone benchmark for video factuality assessment, aiming to steer LVLM development toward verifiable grounding in real-world contexts.

Jing He, Han Zhang, Yuanhui Xiao, Wei Guo, Shaowen Yao, Renyang Liu

Fake news detection methods based on writing style have achieved remarkable progress. However, as adversaries increasingly imitate the style of authentic news, the effectiveness of such approaches is gradually diminishing. Recent research has explored incorporating large language models (LLMs) to enhance fake news detection. Yet, despite their transformative potential, LLMs remain an untapped goldmine for fake news detection, with their real-world adoption hampered by shallow functionality exploration, ambiguous usability, and prohibitive inference costs. In this paper, we propose a novel fake news detection framework, dubbed FACTGUARD, that leverages LLMs to extract event-centric content, thereby reducing the impact of writing style on detection performance. Furthermore, our approach introduces a dynamic usability mechanism that identifies contradictions and ambiguous cases in factual reasoning, adaptively incorporating LLM advice to improve decision reliability. To ensure efficiency and practical deployment, we employ knowledge distillation to derive FACTGUARD-D, enabling the framework to operate effectively in cold-start and resource-constrained scenarios. Comprehensive experiments on two benchmark datasets demonstrate that our approach consistently outperforms existing methods in both robustness and accuracy, effectively addressing the challenges of style sensitivity and LLM usability in fake news detection.

Chengsheng Zhang, Chenghao Sun, Xinyan Jiang, Wei Li, Xinmei Tian

Large Vision-Language Models (LVLMs) have achieved remarkable progress in visual-textual understanding, yet their reliability is critically undermined by hallucinations, i.e., the generation of factually incorrect or inconsistent responses.While recent studies using steering vectors demonstrated promise in reducing hallucinations, a notable challenge remains: they inadvertently amplify the severity of residual hallucinations. We attribute this to their exclusive focus on the decoding stage, where errors accumulate autoregressively and progressively worsen subsequent hallucinatory outputs.To address this, we propose Prefill-Time Intervention (PTI), a novel steering paradigm that intervenes only once during the prefill stage, enhancing the initial Key-Value (KV) cache before error accumulation occurs.Specifically, PTI is modality-aware, deriving distinct directions for visual and textual representations. This intervention is decoupled to steer keys toward visually-grounded objects and values to filter background noise, correcting hallucination-prone representations at their source.Extensive experiments demonstrate PTI's significant performance in mitigating hallucinations and its generalizability across diverse decoding strategies, LVLMs, and benchmarks. Moreover, PTI is orthogonal to existing decoding-stage methods, enabling plug-and-play integration and further boosting performance.

Junha Song, Yongsik Jo, So Yeon Min, Quanting Xie, Taehwan Kim, Yonatan Bisk, Jaegul Choo

Systems such as video chatbots and navigation robots often depend on streaming image captioning to interpret visual inputs. Existing approaches typically employ large multimodal language models (MLLMs) for this purpose, but their substantial computational cost hinders practical application.This limitation motivates our development of a lightweight captioning model. Our investigation begins by replacing the large-scale language component in MLLMs with a compact 125M-parameter model.Surprisingly, this compact model, despite a 93x reduction in size, achieves comparable performance to MLLMs, suggesting that factual image captioning does not significantly require the complex reasoning abilities of LLMs. Despite this promising result, our lightweight model still lacks reliability. To address this, we draw inspiration from the human visual process: perceiving a global and coarse understanding of the scene before attending to finer details. Accordingly, we propose a multimodal self-refinement framework that guides the model to utilize features from salient regions, identified by referencing the previous coarse caption, and to produce a refined description. Experimental results demonstrate the superiority of our model in both single-sentence and detailed captioning, extending even to long-range video QA tasks.