论文检索

Applications · Language, Speech and Dialog

Amirhossein Abaskohi, Tianyi Chen, Miguel Muñoz-Mármol, Curtis Fox, Amrutha Varshini Ramesh, Étienne Marcotte, Xing Han Lu, Nicolas Chapados, Spandana Gella, Christopher Pal 等

We introduce DRBench, a benchmark for evaluating AI agents on complex, open-ended deep research tasks in enterprise settings. Unlike prior benchmarks that focus on simple questions or web-only queries, DRBench evaluates agents on multi-step queries (for example, "What changes should we make to our product roadmap to ensure compliance with this standard?") that require identifying supporting facts from both the public web and private company knowledge base. Each task is grounded in realistic user personas and enterprise context, spanning a heterogeneous search space that includes productivity software, cloud file systems, emails, chat conversations, and the open web. Tasks are generated through a carefully designed synthesis pipeline with human-in-the-loop verification, and agents are evaluated on their ability to recall relevant insights, maintain factual accuracy, and produce coherent, well-structured reports. We release 100 deep research tasks across 10 domains, such as Sales, Cybersecurity, and Compliance. We demonstrate the effectiveness of DRBench by evaluating diverse DR agents across open- and closed-source models (such as GPT, Llama, and Qwen) and DR strategies, highlighting their strengths, weaknesses, and the critical path for advancing enterprise deep research. Code and data are available at https://github.com/ServiceNow/drbench.

Lei You, Yijun Bian, Lele Cao

Counterfactual explanations (CE) aim to reveal how small input changes flip a model’s prediction, yet many methods modify more features than necessary, reducing clarity and actionability. We introduce COLA, a model- and generator-agnostic post-hoc framework that refines any given CE by computing a coupling via optimal transport (OT) between factual and counterfactual sets and using it to drive a Shapley-based attribution p-SHAP that selects a minimal set of edits while preserving the target effect. Theoretically, OT minimizes an upper bound on the $W_1$ divergence between factual and counterfactual outcomes and that, under mild conditions, refined counterfactuals are guaranteed not to move farther from the factuals than the originals. Empirically, across four datasets, twelve models, and five CE generators, COLA achieves the same target effects with only 26–45% of the original feature edits. On a small-scale benchmark, COLA shows near-optimality.

General Machine Learning · Causality

Yuguang Yan, Haolin Yang, Shihao Zhang, Weilin Chen, Ruichu Cai, Zhifeng Hao

In heterogeneous treatment effect estimation from observational data, the fundamental challenge is that only the factual outcome under the received treatment is observable, while the potential outcomes under other treatments or no treatment can never be observed. As a simple and effective approach, matching aims to predict counterfactual outcomes of the target treatment by leveraging the nearest neighbors within the target group. However, due to limited observational data and the distribution shifts between groups, one cannot always find sufficiently close neighbors in the target group, resulting in inaccurate counterfactual prediction because of the manifold structure of data. To address this, we remove group barriers and propose a matching method that selects neighbors from all samples, not just the target group. This helps find closer neighbors and improves counterfactual prediction. Specifically, we analyze the effect estimation error in matching, which motivates us to propose a self optimal transport model for matching. Based on this, we employ an outcome propagation mechanism via the transport plan for counterfactual prediction, and exploit factual outcomes to learn a distance as the transport cost. The experiments are conducted on both binary and multiple treatment settings to evaluate our method.

Deep Learning · Algorithms

Guanzheng Chen, Michael Qizhe Shieh, Lidong Bing

Reinforcement Learning with Verifiable Rewards (RLVR) has significantly advanced the reasoning capabilities of Large Language Models (LLMs) by optimizing them against factual outcomes. However, this paradigm falters in long-context scenarios, as its reliance on internal parametric knowledge is ill-suited for tasks requiring contextual grounding—the ability to find and reason over externally provided information. We identify a key reason for this failure: a reward based solely on the final answer is too sparse to effectively guide the model for identifying relevant evidence. We formally prove that the outcome-only reward leads to significant vanishing gradients for the context grounding process, rendering learning intractable. To overcome this bottleneck, we introduce LongRLVR to augment the sparse answer reward with a dense and verifiable context reward. This auxiliary signal directly incentivizes the model for selecting the correct grounding information, providing a robust learning gradient that solves the underlying optimization challenge. We validate our method on challenging long-context benchmarks using Qwen and LLaMA models. LongRLVR consistently and significantly outperforms the standard RLVR across all models and benchmarks, e.g., boosting a 14B model's scores on RULER-QA from 73.17 to 88.90 and on LongBench v2 from 39.8 to 46.5. Our work demonstrates that explicitly rewarding the grounding process is a critical and effective strategy for unlocking the full reasoning potential of LLMs in long-context applications. Our code is available at https://github.com/real-absolute-AI/LongRLVR.

Xinyu Fang, Zhijian Chen, Kai Lan, Lixin Ma, Shengyuan Ding, Yingji Liang, Xiangyu Zhao, Farong Wen, Zicheng Zhang, Guofeng Zhang 等

Creativity is a fundamental aspect of intelligence, involving the ability to generate novel and appropriate solutions across diverse contexts. While Large Language Models (LLMs) have been extensively evaluated for their creative capabilities, the assessment of Multimodal Large Language Models (MLLMs) in this domain remains largely unexplored. To address this gap, we introduce Creation-MMBench, a multimodal benchmark specifically designed to evaluate the creative capabilities of MLLMs in real-world, image-based tasks. The benchmark comprises 765 test cases spanning 51 fine-grained tasks.To ensure rigorous evaluation, we define instance-specific evaluation criteria for each test case, guiding the assessment of both general response quality and factual consistency with visual inputs. Experimental results reveal that current open-source MLLMs significantly underperform compared to proprietary models in creative tasks. Furthermore, our analysis demonstrates that visual fine-tuning can negatively impact the base LLM's creative abilities.Creation-MMBench provides valuable insights for advancing MLLM creativity and establishes a foundation for future improvements in multimodal generative intelligence. Full data and evaluation code will be released soon.

Lingyong Fang, Xinzhong Wang, Depeng Wang, Zongru Wu, Ya Guo, Huijia Zhu, Zhuosheng Zhang, Gongshen Liu

Multimodal Large Language Models (MLLMs) contain a substantial amount of factual knowledge, which may become outdated or inaccurate over time. Consequently, various knowledge editing techniques have been proposed to update the knowledge encoded within these models. Previous approaches maintain modality consistency during both the editing and testing phases. However, in practical applications, it is desirable for knowledge to be transferable across different modalities, which can enhance the robustness of knowledge editing and potentially allow for cost-effective editing of multimodal knowledge using textual information. To address this, we introduce the concept of Transitivity of Multimodal Knowledge Editing (TMKE) and design corresponding evaluation criteria. Subsequently, we construct a corresponding TMKE Benchmark through an automated pipeline. We evaluate three MLLMs and five knowledge editing methods, uncovering limitations in the current models and methods concerning transitivity. Additionally, we analyze the intrinsic representations of the model during the editing process based on Knowledge Neurons to interpret the experimental phenomena.

Xianfu Cheng, Wei Zhang, Shiwei Zhang, Jian Yang, Xiangyuan Guan, Xianjie Wu, Xiang Li, Ge Zhang, Jiaheng Liu, Yuying Mai 等

The increasing application of multi-modal large language models (MLLMs) across various sectors has spotlighted the essence of their output reliability and accuracy, particularly their ability to produce content grounded in factual information (e.g. common and domain-specific knowledge). In this work, we introduce SimpleVQA, the first comprehensive multi-modal benchmark to evaluate the factuality ability of MLLMs to answer natural language short questions. SimpleVQA is characterized by 7 key features: it is based on bilingual, it covers multiple tasks and multiple scenarios, ensures high quality and challenging queries, maintains static and timeless reference answers, and is straightforward to evaluate. Our approach involves categorizing visual question-answering items into 9 different tasks around objective events or common knowledge and situating these within 9 scenario domains. Rigorous quality control processes are implemented to guarantee high-quality, concise, and clear answers, facilitating evaluation with minimal variance via an LLM-as-a-judge scoring system. Using SimpleVQA, we perform a comprehensive assessment of leading 18 MLLMs and 8 text-only LLMs, delving into their image comprehension and text generation abilities by identifying and analyzing error cases.

Jiaer Xia, Bingkui Tong, Yuhang Zang, Rui Shao, Kaiyang Zhou

Multimodal Large Language Models (MLLMs) have demonstrated remarkable capabilities in interpreting images using natural language. However, without using large-scale datasets for retraining, these models are difficult to adapt to specialized vision tasks, e.g., chart understanding. This problem is caused by a mismatch between pre-training and downstream datasets: pre-training datasets primarily concentrate on scenes and objects but contain limited information about specialized, non-object images, such as charts and tables. In this paper, we share an interesting finding that training an MLLM with chain-of-thought (CoT) reasoning data can facilitate model adaptation in specialized vision tasks, especially under data-limited regimes. However, we identify a critical issue within CoT data distilled from pre-trained MLLMs, i.e., the data often contains multiple factual errors in the reasoning steps. To address the problem, we propose Grounded Chain-of-Thought (GCoT), a simple bootstrapping-based approach that aims to inject grounding information (i.e., bounding boxes) into CoT data, essentially making the reasoning steps more faithful to input images. We evaluate our approach on five specialized vision tasks, which cover a variety of visual formats including charts, tables, receipts, and reports. The results demonstrate that under data-limited regimes our approach significantly improves upon fine-tuning and distillation.

Young-Jun Lee, Byung-Kwan Lee, Jianshu Zhang, Yechan Hwang, Byungsoo Ko, Han-Gyu Kim, Dongyu Yao, Xuankun Rong, Eojin Joo, Seung-Ho Han 等

Vision-and-Language Models (VLMs) have shown impressive capabilities on single-turn benchmarks, yet real-world applications often demand more intricate multi-turn dialogues. Existing multi-turn datasets (e.g, MMDU, ConvBench) only partially capture the breadth and depth of conversational scenarios encountered by users. In this work, we introduce MultiVerse, a novel multi-turn conversation benchmark featuring 647 dialogues--each averaging four turns--derived from a diverse set of 12 popular VLM evaluation benchmarks. With 484 tasks and 484 interaction goals, MultiVerse covers a wide range of topics, from factual knowledge and perception to advanced reasoning tasks such as mathematics and coding. To facilitate robust assessment, we propose a checklist-based evaluation method that leverages GPT-4o as the automated evaluator, measuring performance across 37 key aspects, including perceptual accuracy, linguistic clarity, and factual correctness. We evaluate 18 VLMs on MultiVerse, revealing that even the strongest models (e.g., GPT-4o) achieve only a 50% success rate in complex multi-turn conversations, highlighting the dataset's challenging nature. Notably, we find that providing full dialogue context significantly enhances performance for smaller or weaker models, emphasizing the importance of in-context learning. We believe MultiVerse is a landscape of evaluating multi-turn interaction abilities for VLMs.

Zhongyu Yang, Jun Chen, Dannong Xu, Junjie Fei, Xiaoqian Shen, Liangbing Zhao, Chun-Mei Feng, Mohamed Elhoseiny

Knowledge discovery and collection are intelligence-intensive tasks that traditionally require significant human effort to ensure high-quality outputs. Recent research has explored multi-agent frameworks for automating Wikipedia-style article generation by retrieving and synthesizing information from the internet. However, these methods primarily focus on text-only generation, overlooking the importance of multimodal content in enhancing informativeness and engagement. In this work, we introduce WikiAutoGen, a novel system for automated multimodal Wikipedia-style article generation. Unlike prior approaches, WikiAutoGen retrieves and integrates relevant images alongside text, enriching both the depth and visual appeal of generated content. To further improve factual accuracy and comprehensiveness, we propose a multi-perspective self-reflection mechanism, which critically assesses retrieved content from diverse viewpoints to enhance reliability, breadth, and coherence, etc. Additionally, we introduce WikiSeek, a benchmark comprising Wikipedia articles with topics paired with both textual and image-based representations, designed to evaluate multimodal knowledge generation on more challenging topics. Experimental results show that WikiAutoGen outperforms previous methods by 8%-29% on our WikiSeek benchmark, producing more accurate, coherent, and visually enriched Wikipedia-style articles. We show some of our generated examples in https://anonymous.4open.science/r/WikiAutoGen-C3C4

Mingfeng Zha, Tianyu Li, Guoqing Wang, Peng Wang, Yangyang Wu, Yang Yang, Heng Tao Shen

Audio-visual segmentation (AVS) aims to segment objects in videos based on audio cues. Existing AVS methods are primarily designed to enhance interaction efficiency but pay limited attention to modality representation discrepancies and imbalances. To overcome this, we propose the implicit counterfactual framework (ICF) to achieve unbiased cross-modal understanding. Due to the lack of semantics, heterogeneous representations may lead to erroneous matches, especially in complex scenes with ambiguous visual content or interference from multiple audio sources. We introduce the multi-granularity implicit text (MIT) involving video-, segment- and frame-level as the bridge to establish the modality-shared space, reducing modality gaps and providing prior guidance. Visual content carries more information and typically dominates, thereby marginalizing audio features in the decision-making. To mitigate knowledge preference, we propose the semantic counterfactual (SC) to learn orthogonal representations in the latent space, generating diverse counterfactual samples, thus avoiding biases introduced by complex functional designs and explicit modifications of text structures or attributes. We further formulate the collaborative distribution-aware contrastive learning (CDCL), incorporating factual-counterfactual and inter-modality contrasts to align representations, promoting cohesion and decoupling. Extensive experiments on three public datasets validate that the proposed method achieves state-of-the-art performance.

Thinh-Phuc Nguyen, Thanh-Hai Nguyen, Gia-Huy Dinh, Lam-Huy Nguyen, Minh-Triet Tran, Trung-Nghia Le

Image captioning systems often produce generic descriptions that fail to capture event-level semantics which are crucial for applications like news reporting and digital archiving. We present ReCap, a novel pipeline for event-enriched image retrieval and captioning that incorporates broader contextual information from relevant articles to generate narrative-rich, factually grounded captions. Our approach addresses the limitations of standard vision-language models that typically focus on visible content while missing temporal, social, and historical contexts. ReCap comprises three integrated components: (1) a robust two-stage article retrieval system using DINOv2 embeddings with global feature similarity for initial candidate selection followed by patch-level mutual nearest neighbor similarity re-ranking; (2) a context extraction framework that synthesizes information from article summaries, generic captions, and original source metadata; and (3) a large language model-based caption generation system with Semantic Gaussian Normalization to enhance fluency and relevance. Evaluated on the OpenEvents V1 dataset as part of Track 1 in the EVENTA 2025 Grand Challenge, ReCap achieved a strong overall score of 0.54666, ranking 2nd on the private test set. These results highlight ReCap's effectiveness in bridging visual perception with real-world knowledge, offering a practical solution for context-aware image understanding in high-stakes domains. The code is available at https://github.com/Noridom1/EVENTA2025-Event-Enriched-Image-Captioning.

Shuoping Yang, Jun Yu 0001

We present a unified vision-language framework for the Responsible Multimodal AI Challenge 2025, tackling two related tasks: multimodal hallucination detection and factuality verification. Our approach builds on a Dynamic ViLBERT model enhanced with adaptive multi-branch attention fusion to effectively integrate visual and textual data. For Task A (Hallucination Detection), we employ four parallel answer-encoding branches with co-attentional transformers to compare AI-generated captions or answers against corresponding images, enabling accurate hallucination identification. For Task B (Factuality Verification), we fuse visual and textual features via element-wise addition, multiplication, and concatenation, feeding them into a classifier to assess claim validity. Inputs include visual features from Detectron2 and text embeddings from BERT tokenization. Experiments show notable accuracy gains over baselines, validated by official F1 scores. Extensive ablations, comparative studies, and architectural visualizations confirm the value of cross-modal attention and customized fusion strategies. Our final system ranks 3rd in hallucination detection (F1 = 0.80) and 2nd in factuality verification (F1 = 0.84), demonstrating the strength of our unified approach.

Zijian Zhang 0008, Xuecheng Wu, Danlei Huang, Siyu Yan, Chong Peng, Xuezhi Cao

Driven by the rapid progress in vision-language models (VLMs), the responsible behavior of large-scale multimodal models has become a prominent research area, particularly focusing on hallucination detection and factuality checking. In this paper, we present the solution for the two tracks of Responsible AI challenge. Inspirations from the general domain demonstrate that a smaller distilled VLM can often outperform a larger VLM that is directly tuned on the downstream tasks, while achieving higher efficiency. We thus jointly tackle two tasks from the perspective of knowledge distillation and propose a progressive hybrid knowledge distillation framework termed HKD4VLM. Specifically, the overall framework can be decomposed into Pyramid-like Progressive Online Distillation and Ternary-Coupled Refinement Distillation, hierarchically moving from coarse-grained knowledge alignment to fine-grained refinement. Besides, we further introduce the mapping shift-enhanced inference and diverse augmentation strategies to enhance model performance and robustness. Extensive experimental results demonstrate the effectiveness of our HKD4VLM. Ablation studies provide insights into the critical design choices driving performance gains.

Hoang Chu, Huy Chu, Tan-Minh Nguyen, Son T. Luu, Cuong Hoang, Hiep Nguyen, Vu Tran, Le Minh Nguyen 0001

Significant advancements have been achieved in both fields of Natural Language Processing (NLP) and Computer Vision (CV) with the advent of Multimodal Large Language Models (MLLMs), sometimes referred to as large vision-language models (LVMs). MLLMs show promising ability in multimodal tasks, such as image captioning, visual question answering, etc. However, there is a concerning trend associated with the advancement in MLLMs. These models exhibit an inclination to generate hallucinations and misleading facts, resulting in seemingly plausible yet factually spurious content. To address these challenges, our team, DeepSIX, leverages recent advances in MLLMs to enhance the ability to detect hallucination and verify factual information within the scope of the ACM MM 2025 grand challenge 8: Truthful and Responsible Multimodal Learning (ResMM). We participated in both tasks: Multimodal Hallucination Detection (Task 1) and Multimodal Fact Checking (Task 2). Our approach leverages the interpretive power of the vision and language components of vision language models (VLMs) to analyze and summarize insights from text and images. It performs contextual reasoning by uncovering semantic relationships among entities in the text and objects in the images. By employing diverse prompting techniques, our method deconstructs critical entities in the text, effectively uncovers implicit relationships between text and images, and identifies hallucinations and false facts. Experimental results demonstrate the strength of our approach: it achieved second place in the Hallucination Detection task and third place in the Fact Verification task, confirming the potential of LLM-based methods in MLLMs. We open-source our code at https://github.com/JAIST-DeepSIX/ACMMM25

Xudong Han, Kai Liu 0023, Yanlin Li 0014, Hao Li 0093, Zheng Wang 0007

The Truthful and Responsible Multimodal Learning Challenge aims to foster advancements in the development of reliable and trustworthy multimodal AI systems by addressing two crucial tasks: multimodal hallucination detection and multimodal factuality detection. Task A focuses on detecting hallucinated elements in AI-generated image captions, such as fabricated objects or attributes. Task B targets verifying the factual accuracy of textual claims using visual and contextual cues. We establish benchmarks that support responsible multimodal AI in diverse real-world applications.

Haoxuan Li, Wei Song, Aofan Liu, Peiwu Qin

Document Visual Question Answering (Document VQA) faces significant challenges when processing long documents in low-resource environments due to context limitations and insufficient training data. This paper presents AdaDocVQA, a unified adaptive framework addressing these challenges through three core innovations: a hybrid text retrieval architecture for effective document segmentation, an intelligent data augmentation pipeline that automatically generates high-quality reasoning question-answer pairs with multi-level verification, and adaptive ensemble inference with dynamic configuration generation and early stopping mechanisms. Experiments on Japanese document VQA benchmarks demonstrate substantial improvements with 83.04% accuracy on Yes/No questions, 52.66% on factual questions, and 44.12% on numerical questions in JDocQA, and 59% accuracy on LAVA dataset. Ablation studies confirm meaningful contributions from each component, and our framework establishes new state-of-the-art results for Japanese document VQA while providing a scalable foundation for other low-resource languages and specialized domains. Our code available at: https://github.com/Haoxuanli-Thu/AdaDocVQA.

Zhifei Xie, Hu Zongzheng, Guibin Zhang, Jialin Zhang, Yue Liao, Chunyan Miao, Shuicheng Yan

We present Pask, a proactive AI agent that provides real-time, context-aware guidance and knowledge support in audio-centric media environments. Unlike passive assistants that follow the ''you ask, I answer'' model, Pask shifts toward ''answering before asking'' by continuously monitoring live audio, anticipating user needs, and proactively offering conceptual explanations and semantic clarifications. It integrates three core components: a silent copilot for in-situ explanation, a structured knowledge base for factual grounding, and a private memory module for personalized adaptation. Pask enhances comprehension and communication in scenarios such as online learning, media consumption, and live meetings through sustained, intelligent guidance. A live demo is available at https://www.youtube.com/watch?v=ki_CKiV9Oyk.

Bei Yan, Zhiyuan Chen, Yuecong Min, Jie Zhang 0071, Jiahao Wang, Xiaozhen Wang, Shiguang Shan

Despite rapid advances, Large Vision-Language Models (LVLMs) still suffer from hallucinations, i.e., generating content inconsistent with input or established world knowledge, which correspond to faithfulness and factuality hallucinations, respectively. Prior studies primarily evaluate faithfulness hallucination at a rather coarse level (e.g., object-level) and lack fine-grained analysis. Additionally, existing benchmarks often rely on costly manual curation or reused public datasets, raising concerns about scalability and data leakage. To address these limitations, we propose an automated data construction pipeline that produces scalable, controllable, and diverse evaluation data. We also design a hierarchical hallucination induction framework with input perturbations to simulate realistic noisy scenarios. Integrating these designs, we construct SHALE, a Scalable HALlucination Evaluation benchmark designed to assess both faithfulness and factuality hallucinations via a fine-grained hallucination categorization scheme. SHALE comprises over 30K image-instruction pairs spanning 12 representative visual perception aspects for faithfulness and 6 knowledge domains for factuality, considering both clean and noisy scenarios. Extensive experiments on over 20 mainstream LVLMs reveal significant factuality hallucinations and high sensitivity to semantic perturbations.

Sarmistha Das 0001, R. E. Zera Marveen Lyngkhoi, Sriparna Saha 0001, Alka Maurya

The dynamic propagation of social media has broadened the reach of financial advisory content through podcast videos, yet extracting insights from lengthy, multimodal segments (30-40 minutes) remains challenging. We introduce FASTER(Financial Advisory Summariser with Textual Embedded Relevant images), a modular framework that tackles three key challenges: (1) extracting modality-specific features, (2) producing optimized, concise summaries, and (3) aligning visual keyframes with associated textual points. FASTER employs BLIP-2 for semantic visual descriptions, OCR for textual patterns, and Whisper-based transcription with Speaker diarization as BOS features. A modified Direct Preference Optimization (DPO)-based loss function, equipped with BOS-specific fact-checking, ensures precision, relevance, and factual consistency against the human-aligned summary. A ranker-based retrieval mechanism further aligns keyframes with summarized content, enhancing interpretability and cross-modal coherence. To acknowledge data resource scarcity, we introduce Fin-APT, a dataset comprising 470 publicly accessible financial advisory pep-talk videos for robust multimodal research. Comprehensive cross-domain experiments confirm FASTER's strong performance, robustness, and generalizability when compared to Large Language Models (LLMs) and Vision-Language Models (VLMs). By establishing a new standard for multimodal summarization, FASTER makes financial advisory content more accessible and actionable, thereby opening new avenues for research.