United States

Multimodal Large Language Models (MLLMs) demonstrate remarkable capabilities but often struggle with complex, multi-step mathematical reasoning, where minor errors in visual perception or logical deduction can lead to complete failure. While Process Reward Models (PRMs) offer step-by-step supervision, existing multimodal PRMs are limited to being binary verifiers that can identify but not correct errors, offering little explanatory power. To address these deficiencies, we introduce the Generative Multimodal Process Reward Model (GM-PRM), a novel paradigm that transforms the PRM from a passive judge into an active reasoning collaborator. Instead of a simple scalar score, GM-PRM provides a fine-grained, interpretable analysis of each reasoning step, evaluating its step intent, visual alignment, and logical soundness. More critically, GM-PRM is trained to generate a corrected version of the first erroneous step it identifies. This unique corrective capability enables our new test-time inference strategy, Refined Best-of-N (Refined-BoN). This framework actively enhances solution quality by using the PRM’s generated correction to guide the policy model toward a more promising reasoning trajectory, thereby improving the diversity and correctness of the solution pool. We demonstrate that GM-PRM achieves state-of-the-art results on multiple multimodal math benchmarks, significantly boosting policy model performance with remarkable data efficiency, requiring only a 20K-sample training dataset.

ACL 2026

GM-PRM: A Generative Multimodal Process Reward Model for Multimodal Mathematical Reasoning

process reward model

mathematical reasoning

multimodal learning

## Welcome!
"I am excited to welcome you to this year’s edition of the Conference on Empirical Methods in Natural Language Processing! Importantly, it marks the 30th edition of EMNLP. With over 8,000 submissions, more than 3,000 accepted papers, and thousands of attendees, we have come a long way from that first
workshop, which had 14 accepted papers. As the field looks ahead, Suzhou is the fitting location for celebrating this milestone: rooted in a long literary tradition, yet modern and forward-looking, and home to a large share of the NLP community." 

*Message from the General Chair, Dirk Hovy*

EMNLP 2025

Celebrate 30 Years of EMNLP! 
EMNLP 2025 will be held in Suzhou, China from November 5th to November 9th, 2025.

workshop paper

It is my great pleasure and honor to welcome you to the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026), held in San Diego, California, USA, July 2–7, 2026. ACL 2026 continues our fields tradition of excellence in scholarship, innovation, and inclusivity, and I am deeply grateful to the many volunteers who have worked tirelessly to bring this event to life.

It is exciting to see the many developments in the field reflected at our conference, such as the advancement of large language models (born out of our community) into many application directions. This is not without challenges, partly due to unsolved problems, and partly due to misuse of the technology. All this is reflected well at ACL 2026. The conference also reflects that the field is not only about advancing technology, but also the science of fundamentally understanding language and how it is use by many different speaker communities. 
[Full Message from General Chair Philipp Koehn](https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing) 
<html><button style="display: inline-flex; align-items: center; justify-content: center; white-space: nowrap; border-radius: 9999px; font-weight: bold; background: #7c3aed; color: white; font-family: 'Space Grotesk', sans-serif; height: 40px; font-size: 16px; padding: 0 20px; border: none; cursor: pointer" onclick="window.open('https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing')">Conference Handbook</button></html>

To access this event page, you need to log in with the **email address you registered with**. Access credentials will be sent to your email from Underline - subject line "Welcome to ACL 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://2026.aclweb.org/registration/) first.

**Online Registration Form**: https://acl.swoogo.com/acl2026

Registration Required

The 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026) will take place in San Diego, California, from July 2 - 7, 2026.

Visual grounding in graphical user interface (GUI) requires accurate localization of UI elements from natural language instructions. Conventional coordinate generation approaches face inherent limitations, including sensitivity to resolution variations and lack of interpretability. Recently, coordinate-free attention-based methods have emerged as a promising alternative, but these methods supervise attention using only spatial location signals from ground-truth bounding boxes, without ensuring that the learned attention distributions reflect genuine semantic correspondence between the instruction and the attended visual regions. We propose Attention Cycle-Consistency (ACC), a self-supervised regularization framework that enforces bidirectional alignment between visual attention and instruction semantics. ACC introduces two complementary constraints: semantic consistency, which ensures attended visual regions contain sufficient information to reconstruct the original instruction, and spatial consistency, which requires attention distributions to remain invariant when cycled through instruction reconstruction. We further incorporate entropy regularization to encourage spatially concentrated attention. ACC is applicable as a lightweight, model-agnostic regularizer for attention-based coordinate-free grounding methods, adding zero computational overhead at inference as all auxiliary components are discarded after training.

Look Where You’re Told: Instruction-Consistent Attention for GUI Grounding

Vision–language models (VLMs) achieve strong multimodal performance but remain prone to relation hallucination, which requires accurate reasoning over inter-object interactions. We study the impact of visual perturbations, specifically rotation and noise, and show that even mild distortions significantly degrade relational reasoning across models and datasets. We further evaluate prompt-based augmentation and preprocessing strategies (orientation correction and denoising), finding that while they offer partial improvements, they do not fully resolve hallucinations. Our results reveal a gap between perceptual robustness and relational understanding, highlighting the need for more robust, geometry-aware VLMs.

When Relations Break: Analyzing Relation Hallucination in Vision-Language Model Under Rotation and Noise

General-purpose vision-language models (VLMs) such as LLaVA and QwenVL produce descriptions of disaster imagery that lack domain-specific vocabulary and actionable detail. We propose the Vision-Language Caption Enhancer (VLCE), a framework that integrates external semantic knowledge from ConceptNet and WordNet into the caption generation process for post-disaster satellite and UAV imagery. VLCE operates in two stages: first, a baseline VLM generates an initial caption conditioned on YOLOv8 object detections; second, a knowledge-enriched sequential model, a CNN-LSTM or a hierarchical cross-modal Transformer, refines the caption using a vocabulary augmented with 1,566 domain-relevant terms extracted from knowledge graphs. We evaluate VLCE on two disaster benchmarks: xBD (satellite, 6,369 images, 3 damage classes) and RescueNet (UAV, 4,494 images, 12 damage classes), using CLIPScore for semantic alignment and InfoMetIC for informativeness. On RescueNet with the Transformer decoder, VLCE with knowledge graph enrichment produces captions preferred over QwenVL baselines in 95.33% of image pairs on InfoMetIC and 73.64% on CLIPScore. Qualitative analysis shows that without knowledge graph integration, generated captions exhibit hallucinations, word repetition, and semantic incoherence, whereas knowledge-enriched captions maintain factual consistency and domain-appropriate vocabulary. intended as a continuous, extensible monitor of differential framing under changing real-world inputs.

VLCE: A Knowledge-Enhanced Framework for Image Description in Disaster Assessment

Traditional clustering techniques often rely solely on similarity in the input data, limiting their ability to capture structural or semantic constraints that are critical in many domains. We introduce the Domain-Aware Rule-Triggered Variational Autoencoder (DART-VAE), a rule-guided multimodal clustering framework that incorporates domain-specific constraints directly into the representation learning process. DART-VAE extends the VAE architecture by embedding explicit rules, semantic representations, and data-driven features into a unified latent space, while enforcing constraint compliance through rule-consistency and violation penalties in the loss function. Unlike conventional clustering methods that rely only on visual similarity or apply rules as post-hoc filters, DART-VAE treats rules as first-class learning signals. The rules are generated by LLMs, structured into knowledge graphs, and enforced through a loss function combining reconstruction, KL divergence, consistency, and violation penalties. Experiments on aircraft and automotive datasets demonstrate that rule-guided clustering produces more operationally meaningful and interpretable clusters—for example, isolating UAVs, unifying stealth aircraft, or separating SUVs from sedans—while improving traditional clustering metrics. However, the framework faces challenges: LLM-generated rules may hallucinate or conflict, excessive rules risk overfitting, and scaling to complex domains increases computational and consistency difficulties. By combining rule encodings with learned representations, DART-VAE achieves more meaningful and consistent clustering outcomes than purely data-driven models, highlighting the utility of constraint-guided multimodal clustering for complex, knowledge-intensive settings.

Beyond Visual Similarity: Rule-Guided Multimodal Clustering with explicit domain rules

ChartDiff is a dataset and benchmark for cross-chart comparative summarization. It contains 8,541 chart pairs with high-quality natural-language comparison annotations. The charts cover six chart types: line, bar, horizontal bar, multi-series line, multi-series bar, and pie. The rendering uses Matplotlib, Plotly, and Plotnine. The data sources include Macrotrends, Yahoo Finance, and Visual Crossing. The annotation pipeline is annotate-judge-verify. The benchmark evaluates general-purpose VLMs, chart-specialized models, and pipeline-based methods.

ChartDiff: A Large-Scale Benchmark for Comprehending Pairs of Charts

The precolonial and early colonial Mixtec codices describe the history and stories of the region in a semasiographic medium that is full of symbolic representations and meant to be narrated. Recently, the community has investigated hierarchical representation of related media, including Aztec codices and Mayan hieroglyphic script, in a step towards symbolic machine interpretation of these historic Mesoamerican artifacts. In this work, we propose formal symbolic machine interpretation of XML encodings representing facsimile images from the Mixtec Codex Zouche-Nuttall. We demonstrate the efficacy of symbolic machine interpretation from XML step-by-step, showing how our parser and interpreter process text capturing a scene from the Mixtec Codex Zouche-Nuttall. We hope our contribution and the example we provide motivate collaboration among the archaeological, historical, linguistic, and natural language processing research communities to apply machine interpretation to Mixtec codices and similar manuscripts.

Formal Machine Interpretation for the Semasiographic Mixtec Codices of Precolonial and Early Colonial Mesoamerica

One of the expected abilities of vision-language models (VLMs) is spatial reasoning ability based on a given text and image. To evaluate the spatial reasoning abilities of VLMs, we focus on the use of spatial deictic expressions, which are defined as spatial expressions whose referent is determined by their situational context, such as this and that. To handle spatial deictic expressions, VLMs must jointly reason over language and visual space, grounding context-dependent references in the image's spatial structure. In addition, selecting appropriate spatial deictic expressions across languages requires VLMs to understand the language-specific spatial distinctions encoded by these expressions. In this paper, we develop a benchmark to evaluate the multilingual ability of VLMs to use spatial deictic expressions in four languages. Our experiments using this benchmark reveal that the tested models use demonstratives in a manner different from that of humans, particularly in selecting the appropriate demonstratives based on the distance from the object.

Evaluation of Multilingual Ability to Use Spatial Deictic Expressions in Vision-Language Models

Multimodal Large Language Models (MLLMs) have achieved remarkable success in semantic visual reasoning, yet their capacity for fine-grained, low-level perception remains critically under-evaluated. This perceptual fragility limits their reliability in noisy, real-world environments where visual signals are degraded. Furthermore, existing benchmarks often entangle visual perception with language priors, masking these underlying deficits. To address this, we introduce the \textbf{FAint numeric Detection Evaluation (FADE)} dataset, a novel evaluation suite designed to probe the limits of zero-shot Optical Character Recognition (OCR) in frontier MLLMs. By embedding synthetic, strictly numerical sequences over cluttered natural backgrounds at varying levels of transparency ($\alpha$), FADE explicitly disentangles pure visual perception from semantic predictability. We evaluate state-of-the-art models including Gemini 3.0, Claude 4.5 Sonnet, and Gemma 3 against a specialized UNet segmentation baseline. Our results reveal a striking limitation in frontier architectures: while they achieve near-perfect transcription at high visibility, their performance collapses under high transparency. Conversely, the UNet pipeline maintains robust spatial grounding, significantly outperforming generalist models at the lowest visibility thresholds. FADE provides a reproducible dataset to expose and diagnose the perceptual breakage points of modern multimodal systems.

FADE: Probing the Limits of VLMs on fine-grained OCR

We propose Question-Guided Sparse Attention (QGSA), a lightweight plug-and-play module that reduces visual redundancy in VQA by dynamically selecting only question-relevant image patches before cross-attention. Using a Gumbel-Softmax differentiable selector and a self-supervised grounding loss requiring no bounding-box annotations, QGSA reduces cross-attention FLOPs by 91–99% across standard resolutions while preserving exact accuracy on VQA-RAD, A-OKVQA, and RefCOCO. On SmolVLM-256M, it selects ~17 of 576 patches (34× token compression) and achieves a 31× relative improvement in grounding IoU over random selection. Wall-clock parity with the dense baseline is reached at 336px, with 76.2× theoretical speedup at 576px. These results validate the feasibility of question-conditioned sparse attention as a foundation for efficient inference in larger vision-language models.

Efficient Visual Grounding in VQA via Question-Guided Sparse Attention

Mapping linguistic to visual representations is a central objective in vision-language model (VLM) development. Consequently, current VLM benchmarks and training objectives predominantly optimize for literal, descriptive correspondence between modalities. However, in real-world multimedia contexts, text and images are rarely strictly redundant; they are often loosely coupled, interacting at higher narrative levels. To capture this dynamic, we introduce Visual Narrative Freedom (VNF): the degree to which a generative system can produce multiple plausible visual realizations from an underdetermined textual input. In this paper, we systematically evaluate text-to-image (T2I) generation across a continuum of linguistic constraints, conditioning models on visual descriptions, captions, and full news articles. Using perceptual, structural, and semantic metrics, we first demonstrate that as textual constraints loosen (and VNF increases), T2I models produce images that are significantly more diverse from ground-truth reference photographs. Human evaluation further reveals that providing generation models with greater visual narrative freedom significantly increases the likelihood that their outputs are preferred over authentic news photographs, despite the known stylistic limitations of AI imagery. Ultimately, these findings suggest that the creation of persuasive multimodal misinformation is more imminent than evaluations of T2I systems based solely on text descriptions may indicate. This highlights the need for VLM evaluation frameworks that better capture the underdetermined nature of real-world image-text alignment.

Downloads

Next from ACL 2026

Look Where You’re Told: Instruction-Consistent Attention for GUI Grounding

Similar lecture

R-PRM: Reasoning-Driven Process Reward Modeling

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from ACL 2026

Look Where You’re Told: Instruction-Consistent Attention for GUI Grounding

Similar lecture

R-PRM: Reasoning-Driven Process Reward Modeling

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Downloads