Multimodal Retrieval-Augmented Generation (MRAG) enables Multimodal Large Language Models (MLLMs) to generate responses with external multimodal evidence. Recently, numerous video-based MRAG benchmarks have been proposed to evaluate model capabilities across retrieval and generation stages in MRAG. However, existing benchmarks remain limited in modality coverage and format diversity, often focusing on single- or limited-modality tasks, or coarse-grained scene understanding. To address these gaps, we introduce CFVBench, a large-scale, manually verified benchmark constructed from 599 publicly available videos, yielding 5,360 open-ended QA pairs. CFVBench, spans high-density formats and domains such as chart-heavy reports, news broadcasts, and software tutorials, requiring models to retrieve and reason over long temporal video spans while maintaining fine-grained multimodal information. Using CFVBench, we systematically evaluate 7 retrieval methods and 14 widely-used MLLMs, revealing a critical bottleneck: current models (even GPT5 or Gemini) struggle to capture transient yet essential fine-grained multimodal details. To mitigate this, we propose Adaptive Visual Refinement (AVR), a plug-and-paly framework that adaptively increases frame sampling density and selectively invokes external tools when necessary. Experiments show that AVR consistently enhances fine-grained multimodal comprehension and improves performance across all evaluated MLLMs.

Jie Zhang, Changzai Pan, Sishi Xiong, Kaiwen Wei, Yu Zhao, Xiangyu Li, Jiaxin Peng, Xiaoyan Gu, Jian Yang, Wenhan Chang, Zhenhe Wu, Jiang Zhong, Shuangyong Song, Xuelong Li. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Timely detection and diagnosis of diseases are key elements of an efficient healthcare system. In recent years, artificial intelligence (AI) has played an increasingly important role in improving the accuracy and efficiency of disease diagnosis in clinical practice. However, most existing AI systems for disease diagnosis have focused on either classifying patients into broad disease categories or diagnosing a specific disease, leaving a gap in the development of a coherent AI system for both triage and diagnosis in a department of a general hospital. In this study, we fill this gap with SmartGyne, an advanced AI system that can achieve high-quality triage and diagnosis for a full spectrum of gynecological diseases. By extracting useful clinical evidence for diagnosis from a large amount of electronic medical records, SmartGyne establishes an effective framework to integrate real-world clinical evidence and knowledge into a coherent AI system that can effectively handle a full spectrum of complex diseases in a department of a general hospital. Validation experiments demonstrated that SmartGyne achieved an overall accuracy of 80.1 % in triage for gynecological diseases, and 99.4 % in diagnosis for a gynecological subspecialty. In comparison with human physicians, SmartGyne showed competitive triage and diagnostic performance, and improved consultation efficiency and accuracy for physicians with limited specialized experience. These results show that SmartGyne achieves high-quality triage and diagnosis, holding the potential to improve the efficiency of the healthcare system in China, as well as other countries lacking professional gynecologists.

As a precious cultural heritage of human beings, classical Chinese poetry has a very unique writing style and often contains special words that rarely appear in general Chinese texts, posting critical challenges for natural language processing. Little effort has been made in the literature for processing texts from classical Chinese poetry. This study fills in this gap with TopWORDS-Poetry, an unsupervised method that can achieve reliable text segmentation and word discovery for classical Chinese poetry simultaneously without pre-given vocabulary or training corpus. Experimental studies confirm that TopWORDS-Poetry can successfully recognize unique poetry words, such as named entities and literary allusions, from metrical poems of Complete Tang Poetry and segment these poetry lines into sequences of meaningful words with high quality.

Processing open-domain Chinese texts has been a critical bottleneck in computational linguistics for decades, partially because text segmentation and word discovery often entangle with each other in this challenging scenario.No existing methods yet can achieve effective text segmentation and word discovery simultaneously in open domain.This study fills in this gap by proposing a novel method called TopWORDS-Seg based on Bayesian inference, which enjoys robust performance and transparent interpretation when no training corpus and domain vocabulary are available.Advantages of TopWORDS-Seg are demonstrated by a series of experimental studies.