Zhixiong Zeng

Yufeng Zhong, Lei Chen, Xuanle Zhao et al.

The development of large vision language models drives the demand for managing, and applying massive amounts of multimodal data, making OCR technology, which extracts information from visual images, increasingly popular. However, existing OCR methods primarily focus on recognizing text elements from images or scanned documents (Text-centric OCR), neglecting the identification of visual elements from visually information-dense image sources (Vision-centric OCR), such as charts, web pages and science plots. In reality, these visually information-dense images are widespread on the internet and have significant real-world application value, such as data visualization and web page analysis. In this technical report, we propose OCRVerse, the first holistic OCR method in end-to-end manner that enables unified text-centric OCR and vision-centric OCR. To this end, we constructe comprehensive data engineering to cover a wide range of text-centric documents, such as newspapers, magazines and books, as well as vision-centric rendered composites, including charts, web pages and scientific plots. Moreover, we propose a two-stage SFT-RL multi-domain training method for OCRVerse. SFT directly mixes cross-domain data to train and establish initial domain knowledge, while RL focuses on designing personalized reward strategies for the characteristics of each domain. Specifically, since different domains require various output formats and expected outputs, we provide sufficient flexibility in the RL stage to customize flexible reward signals for each domain, thereby improving cross-domain fusion and avoiding data conflicts. Experimental results demonstrate the effectiveness of OCRVerse, achieving competitive results across text-centric and vision-centric data types, even comparable to large-scale open-source and closed-source models.

Xuanle Zhao, Deyang Jiang, Zhixiong Zeng et al.

Multimodal code generation has garnered significant interest within the research community. Despite the notable success of recent vision-language models (VLMs) on specialized tasks like Chart-to-code generation, their reliance on single-task training regimens fosters a narrow paradigm that hinders the development of generalized \textbf{VI}sio\textbf{N} \textbf{C}ode \textbf{I}ntelligence. In this work, we introduce \textbf{VinciCoder}, a unified multimodal code generation model that addresses this limitation via a two-stage training framework. We begin by constructing a large-scale Supervised Finetuning (SFT) corpus comprising 1.6M image-code pairs for tasks involving direct code generation and visual-based code refinement. Subsequently, we introduce a Visual Reinforcement Learning (ViRL) strategy, which employs a coarse-to-fine reward mechanism to improve visual fidelity by calculating visual similarity across local and global image patches. Extensive experiments on various multimodal code generation benchmarks demonstrate that VinciCoder achieves state-of-the-art performance, underscoring the effectiveness of our coarse-to-fine ViRL strategy. The code and model will be available at https://github.com/DocTron-hub/VinciCoder.

Deyang Jiang, Jing Huang, Xuanle Zhao et al.

Effectively scaling GUI automation is essential for computer-use agents (CUAs); however, existing work primarily focuses on scaling GUI grounding rather than the more crucial GUI planning, which requires more sophisticated data collection. In reality, the exploration process of a CUA across apps/desktops/web pages typically follows a tree structure, with earlier functional entry points often being explored more frequently. Thus, organizing large-scale trajectories into tree structures can reduce data cost and streamline the data scaling of GUI planning. In this work, we propose TreeCUA to efficiently scale GUI automation with tree-structured verifiable evolution. We propose a multi-agent collaborative framework to explore the environment, verify actions, summarize trajectories, and evaluate quality to generate high-quality and scalable GUI trajectories. To improve efficiency, we devise a novel tree-based topology to store and replay duplicate exploration nodes, and design an adaptive exploration algorithm to balance the depth (\emph{i.e.}, trajectory difficulty) and breadth (\emph{i.e.}, trajectory diversity). Moreover, we develop world knowledge guidance and global memory backtracking to avoid low-quality generation. Finally, we naturally extend and propose the TreeCUA-DPO method from abundant tree node information, improving GUI planning capability by referring to the branch information of adjacent trajectories. Experimental results show that TreeCUA and TreeCUA-DPO offer significant improvements, and out-of-domain (OOD) studies further demonstrate strong generalization. All trajectory node information and code will be available at https://github.com/UITron-hub/TreeCUA.

Yiming Ju, Xingrun Xing, Zhixiong Zeng

Recently, Locate-Then-Edit paradigm has emerged as one of the main approaches in changing factual knowledge stored in the Language models. However, there is a lack of research on whether present locating methods can pinpoint the exact parameters embedding the desired knowledge. Moreover, although many researchers have questioned the validity of locality hypothesis of factual knowledge, no method is provided to test the a hypothesis for more in-depth discussion and research. Therefore, we introduce KLoB, a benchmark examining three essential properties that a reliable knowledge locating method should satisfy. KLoB can serve as a benchmark for evaluating existing locating methods in language models, and can contributes a method to reassessing the validity of locality hypothesis of factual knowledge. KLoB is publicly available at an anonymous GitHub: \url{https://github.com/anon6662/KLoB}.

Siqi Yang, Zilve Gao, Haibo Qiu et al.

Multimodal Large Language Models (MLLMs) demonstrate significant potential but remain brittle in complex, long-chain visual reasoning tasks. A critical failure mode is "visual forgetting", where models progressively lose visual grounding as reasoning extends, a phenomenon aptly described as "think longer, see less". We posit this failure stems from current training paradigms prematurely entangling two distinct cognitive skills: (1) abstract logical reasoning "how-to-think") and (2) strategic visual perception ("when-to-look"). This creates a foundational cold-start deficiency -- weakening abstract reasoning -- and a strategic perception deficit, as models lack a policy for when to perceive. In this paper, we propose a novel curriculum-based framework to disentangle these skills. First, we introduce a disentangled Supervised Fine-Tuning (SFT) curriculum that builds a robust abstract reasoning backbone on text-only data before anchoring it to vision with a novel Perception-Grounded Chain-of-Thought (PG-CoT) paradigm. Second, we resolve the strategic perception deficit by formulating timing as a reinforcement learning problem. We design a Pivotal Perception Reward that teaches the model when to look by coupling perceptual actions to linguistic markers of cognitive uncertainty (e.g., "wait", "verify"), thereby learning an autonomous grounding policy. Our contributions include the formalization of these two deficiencies and the development of a principled, two-stage framework to address them, transforming the model from a heuristic-driven observer to a strategic, grounded reasoner. \textbf{Code}: \url{https://github.com/gaozilve-max/learning-when-to-look}.

Meituan LongCat Team, Bairui Wang, Bayan et al.

We introduce LongCat-Flash-Omni, a state-of-the-art open-source omni-modal model with 560 billion parameters, excelling at real-time audio-visual interaction. By adopting a curriculum-inspired progressive training strategy that transitions from simpler to increasingly complex modality sequence modeling tasks, LongCat-Flash-Omni attains comprehensive multimodal capabilities while maintaining strong unimodal capability. Building upon LongCat-Flash, which adopts a high-performance Shortcut-connected Mixture-of-Experts (MoE) architecture with zero-computation experts, LongCat-Flash-Omni integrates efficient multimodal perception and speech reconstruction modules. Despite its immense size of 560B parameters (with 27B activated), LongCat-Flash-Omni achieves low-latency real-time audio-visual interaction. For training infrastructure, we developed a modality-decoupled parallelism scheme specifically designed to manage the data and model heterogeneity inherent in large-scale multimodal training. This innovative approach demonstrates exceptional efficiency by sustaining over 90% of the throughput achieved by text-only training. Extensive evaluations show that LongCat-Flash-Omni achieves state-of-the-art performance on omni-modal benchmarks among open-source models. Furthermore, it delivers highly competitive results across a wide range of modality-specific tasks, including text, image, and video understanding, as well as audio understanding and generation. We provide a comprehensive overview of the model architecture design, training procedures, and data strategies, and open-source the model to foster future research and development in the community.

Lei Chen, Xuanle Zhao, Zhixiong Zeng et al.

Recently, inspired by OpenAI-o1/o3 and Deepseek-R1, the R1-Style method based on reinforcement learning fine-tuning has received widespread attention from the community. Previous R1-Style methods mainly focus on mathematical reasoning and code intelligence. It is of great research significance to verify their advantages on more general multimodal data. Chart is an important multimodal data type with rich information, which brings important research challenges in complex reasoning. In this work, we introduce Chart-R1, a chart-domain vision-language model with reinforcement learning fine-tuning to enable complex chart reasoning. To support Chart-R1, we first propose a novel programmatic data synthesis technology to generate high-quality step-by-step chart reasoning data covering single- and multi-subcharts, which makes up for the lack of reasoning data in the chart domain. Then we develop a two-stage training strategy: Chart-COT with step-by-step chain-of-thought supervision, and Chart-RFT with numerically sensitive reinforcement fine-tuning. Chart-COT aims to decompose complex chart reasoning tasks into fine-grained, understandable subtasks through step-by-step supervision, which lays a good foundation for improving the reasoning level of reinforcement learning. Chart-RFT utilize the typical group relative policy optimization strategy, in which a relatively soft reward is adopted for numerical response to emphasize the numerical sensitivity in the chart domain. We conduct extensive experiments on open-source benchmarks and self-built chart reasoning dataset (\emph{i.e., ChartRQA}). Experimental results show that Chart-R1 has significant advantages compared to chart-domain methods, even comparable to open/closed source large-scale models (\emph{e.g., GPT-4o, Claude-3.5}).

Zhixiong Zeng, Jing Huang, Liming Zheng et al.

GUI agent aims to enable automated operations on Mobile/PC devices, which is an important task toward achieving artificial general intelligence. The rapid advancement of VLMs accelerates the development of GUI agents, owing to their powerful capabilities in visual understanding and task planning. However, building a GUI agent remains a challenging task due to the scarcity of operation trajectories, the availability of interactive infrastructure, and the limitation of initial capabilities in foundation models. In this work, we introduce UItron, an open-source foundational model for automatic GUI agents, featuring advanced GUI perception, grounding, and planning capabilities. UItron highlights the necessity of systemic data engineering and interactive infrastructure as foundational components for advancing GUI agent development. It not only systematically studies a series of data engineering strategies to enhance training effects, but also establishes an interactive environment connecting both Mobile and PC devices. In training, UItron adopts supervised finetuning over perception and planning tasks in various GUI scenarios, and then develop a curriculum reinforcement learning framework to enable complex reasoning and exploration for online environments. As a result, UItron achieves superior performance in benchmarks of GUI perception, grounding, and planning. In particular, UItron highlights the interaction proficiency with top-tier Chinese mobile APPs, as we identified a general lack of Chinese capabilities even in state-of-the-art solutions. To this end, we manually collect over one million steps of operation trajectories across the top 100 most popular apps, and build the offline and online agent evaluation environments. Experimental results demonstrate that UItron achieves significant progress in Chinese app scenarios, propelling GUI agents one step closer to real-world application.

Wenkang Han, Zhixiong Zeng, Jing Huang et al.

Autonomous agents for Graphical User Interfaces (GUIs) are revolutionizing human-computer interaction, yet their reliance on text-based instructions imposes limitations on accessibility and convenience, particularly in hands-free scenarios. To address this issue, we propose replacing text with speech as the instruction input modality for GUI agents, and introduce UITron-Speech, which is the first end-to-end GUI agent capable of directly processing speech instructions and on-device screenshots to predict user actions. To tackle the problem of data scarcity, we synthesize high-quality speech instruction datasets using a random-speaker text-to-speech model. Additionally, we design a mixed-modality training strategy to mitigate the inherent modality imbalance in pre-trained foundation models. Furthermore, we conduct a statistical analysis of the distribution of GUI grounding prediction errors and propose a training-free two-step grounding refinement method to alleviate minor localization deviations. Extensive experiments on multiple benchmarks demonstrate that UITron-Speech achieves robust performance and superior adaptability, underscoring the feasibility and potential of speech-driven GUI agents for more accessible and intelligent human-computer interaction. Our code and datasets are available at https://github.com/UITron-hub/UITron-Speech.

Kun Chen, Peng Shi, Fanfan Liu et al.

Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a critical method for enhancing the reasoning capabilities of Large Language Models (LLMs). However, continuous training often leads to policy entropy collapse, characterized by a rapid decay in entropy that results in premature overconfidence, reduced output diversity, and vanishing gradient norms that inhibit learning. Gradient-Preserving Clipping is a primary factor influencing these dynamics, but existing mitigation strategies are largely static and lack a framework connecting clipping mechanisms to precise entropy control. This paper proposes reshaping entropy control in RL from the perspective of Gradient-Preserving Clipping. We first theoretically and empirically verify the contributions of specific importance sampling ratio regions to entropy growth and reduction. Leveraging these findings, we introduce a novel regulation mechanism using dynamic clipping threshold to precisely manage entropy. Furthermore, we design and evaluate dynamic entropy control strategies, including increase-then-decrease, decrease-increase-decrease, and oscillatory decay. Experimental results demonstrate that these strategies effectively mitigate entropy collapse, and achieve superior performance across multiple benchmarks.

Yilin Cao, Yufeng Zhong, Zhixiong Zeng et al.

Mobile GUI agents have shown strong potential in real-world automation and practical applications. However, most existing agents remain reactive, making decisions mainly from current screen, which limits their performance on long-horizon tasks. Building a world model from repeated interactions enables forecasting action outcomes and supports better decision making for mobile GUI agents. This is challenging because the model must predict post-action states with spatial awareness while remaining efficient enough for practical deployment. In this paper, we propose MobileDreamer, an efficient world-model-based lookahead framework to equip the GUI agents based on the future imagination provided by the world model. It consists of textual sketch world model and rollout imagination for GUI agent. Textual sketch world model forecasts post-action states through a learning process to transform digital images into key task-related sketches, and designs a novel order-invariant learning strategy to preserve the spatial information of GUI elements. The rollout imagination strategy for GUI agent optimizes the action-selection process by leveraging the prediction capability of world model. Experiments on Android World show that MobileDreamer achieves state-of-the-art performance and improves task success by 5.25%. World model evaluations further verify that our textual sketch modeling accurately forecasts key GUI elements.

Fanfan Liu, Youyang Yin, Peng Shi et al.

Recent applications of Reinforcement Learning with Verifiable Rewards (RLVR) to Large Language Models (LLMs) and Vision-Language Models (VLMs) have demonstrated significant success in enhancing reasoning capabilities for complex tasks. During RLVR training, an increase in response length is often regarded as a key factor contributing to the growth of reasoning ability. However, the patterns of change in response length vary significantly across different RLVR algorithms during the training process. To provide a fundamental explanation for these variations, this paper conducts an in-depth analysis of the components of mainstream RLVR algorithms. We present a theoretical analysis of the factors influencing response length and validate our theory through extensive experimentation. Building upon these theoretical findings, we propose the Length-Unbiased Sequence Policy Optimization (LUSPO) algorithm. Specifically, we rectify the length bias inherent in Group Sequence Policy Optimization (GSPO), rendering its loss function unbiased with respect to response length and thereby resolving the issue of response length collapse. We conduct extensive experiments across mathematical reasoning benchmarks and multimodal reasoning scenarios, where LUSPO consistently achieves superior performance. Empirical results demonstrate that LUSPO represents a novel, state-of-the-art optimization strategy compared to existing methods such as GRPO and GSPO.

Yanbin Wei, Shuai Fu, Weisen Jiang et al.

Large Language Models (LLMs) are increasingly used for various tasks with graph structures. Though LLMs can process graph information in a textual format, they overlook the rich vision modality, which is an intuitive way for humans to comprehend structural information and conduct general graph reasoning. The potential benefits and capabilities of representing graph structures as visual images (i.e., $\textit{visual graph}$) are still unexplored. To fill the gap, we innovatively propose an end-to-end framework, called $\textbf{G}$raph to v$\textbf{I}$sual and $\textbf{T}$extual Integr$\textbf{A}$tion (GITA), which firstly incorporates visual graphs into general graph reasoning. Besides, we establish $\textbf{G}$raph-based $\textbf{V}$ision-$\textbf{L}$anguage $\textbf{Q}$uestion $\textbf{A}$nswering (GVLQA) dataset from existing graph data, which is the first vision-language dataset for general graph reasoning purposes. Extensive experiments on the GVLQA dataset and five real-world datasets show that GITA outperforms mainstream LLMs in terms of general graph reasoning capabilities. Moreover, We highlight the effectiveness of the layout augmentation on visual graphs and pretraining on the GVLQA dataset.

Jing Huang, Zhixiong Zeng, Wenkang Han et al.

Automated GUI agents aims to facilitate user interaction by automatically performing complex tasks in digital environments, such as web, mobile, desktop devices. It receives textual task instruction and GUI description to generate executable actions (\emph{e.g.}, click) and operation boxes step by step. Training a GUI agent mainly involves grounding and planning stages, in which the GUI grounding focuses on finding the execution coordinates according to the task, while the planning stage aims to predict the next action based on historical actions. However, previous work suffers from the limitations of insufficient training data for GUI grounding, as well as the ignorance of backtracking historical behaviors for GUI planning. To handle the above challenges, we propose ScaleTrack, a training framework by scaling grounding and backtracking planning for automated GUI agents. We carefully collected GUI samples of different synthesis criterions from a wide range of sources, and unified them into the same template for training GUI grounding models. Moreover, we design a novel training strategy that predicts the next action from the current GUI image, while also backtracking the historical actions that led to the GUI image. In this way, ScaleTrack explains the correspondence between GUI images and actions, which effectively describes the evolution rules of the GUI environment. Extensive experimental results demonstrate the effectiveness of ScaleTrack. Data and code will be available at url.

Lei Chen, Xuanle Zhao, Zhixiong Zeng et al.

While reinforcement learning (RL) has proven highly effective for general reasoning in vision-language models, its application to tasks requiring in-depth understanding of information-rich images and generation of structured outputs remains underexplored. Chart-to-code generation exemplifies this challenge, demanding complex reasoning over visual charts to generate structured code. Supervised fine-tuning (SFT) alone is often insufficient, highlighting the need for effective RL strategies that appropriately reward structured outputs. We systematically investigate the performance plateau in SFT through large-scale experiments and propose Multimodal Structured Reinforcement Learning (MSRL) for chart-to-code generation, which substantially breaks through this plateau. We construct the largest training corpus to date, containing 3 million chart-code pairs from real-world arXiv tables to mitigate simplistic patterns of prior synthetic data. Despite reaching state-of-the-art performance, our experiments show that scaling SFT data eventually hits a plateau where further increases yield negligible improvements. Our MSRL method leverages a multi-granularity structured reward system using multimodal textual and visual feedback. At the textual level, rule-based rewards validate fine-grained code details. At the visual level, model-based rewards assess structural similarity by rendering generated code into images and employing an evaluator model. We implement this within a two-stage curriculum for training stability. Results demonstrate that MSRL significantly breaks the SFT plateau, improving high-level metrics by 6.2% and 9.9% on ChartMimic and ReachQA benchmarks respectively, achieving competitive performance with advanced closed-source models.

Chi Zhang, Haibo Qiu, Qiming Zhang et al.

The "thinking with images" paradigm represents a pivotal shift in the reasoning of Vision Language Models (VLMs), moving from text-dominant chain-of-thought to image-interactive reasoning. By invoking visual tools or generating intermediate visual representations, VLMs can iteratively attend to fine-grained regions, enabling deeper image understanding and more faithful multimodal reasoning. As an emerging paradigm, however, it still leaves substantial room for exploration in data construction accuracy, structural design, and broader application scenarios, which offer rich opportunities for advancing multimodal reasoning. To further advance this line of work, we present DeepSketcher, a comprehensive suite comprising both an image-text interleaved dataset and a self-contained model. The dataset contains 31k chain-of-thought (CoT) reasoning trajectories with diverse tool calls and resulting edited images, covering a wide range of data types and manipulation instructions with high annotation accuracy. Building on this resource, we design a model that performs interleaved image-text reasoning and natively generates "visual thoughts" by operating directly in the visual embedding space, rather than invoking external tools and repeatedly re-encoding generated images. This design enables tool-free and more flexible "thinking with images". Extensive experiments on multimodal reasoning benchmarks demonstrate strong performance, validating both the utility of the dataset and the effectiveness of the model design.

Longrong Yang, Zhixiong Zeng, Yufeng Zhong et al.

Multimodal large language models are evolving toward multimodal agents capable of proactively executing tasks. Most agent research focuses on GUI or embodied scenarios, which correspond to agents interacting with 2D virtual worlds or 3D real worlds, respectively. However, many complex tasks typically require agents to interleavely interact with these two types of environment. We initially mix GUI and embodied data to train, but find the performance degeneration brought by the data conflict. Further analysis reveals that GUI and embodied data exhibit synergy and conflict at the shallow and deep layers, respectively, which resembles the cerebrum-cerebellum mechanism in the human brain. To this end, we propose a high-performance generalist agent OmniActor, designed from both structural and data perspectives. First, we propose Layer-heterogeneity MoE to eliminate the conflict between GUI and embodied data by separating deep-layer parameters, while leverage their synergy by sharing shallow-layer parameters. By successfully leveraging the synergy and eliminating the conflict, OmniActor outperforms agents only trained by GUI or embodied data in GUI or embodied tasks. Furthermore, we unify the action spaces of GUI and embodied tasks, and collect large-scale GUI and embodied data from various sources for training. This significantly improves OmniActor under different scenarios, especially in GUI tasks. The code will be publicly available.

Yufeng Zhong, Lei Chen, Zhixiong Zeng et al.

Reading text from images or scanned documents via OCR models has been a longstanding focus of researchers. Intuitively, text reading is perceived as a straightforward perceptual task, and existing work primarily focuses on constructing enriched data engineering to enhance SFT capabilities. In this work, we observe that even advanced OCR models exhibit significantly higher entropy in formatted text (\emph{e.g.}, formula, table, etc.) compared to plain text, often by an order of magnitude. These statistical patterns reveal that advanced OCR models struggle with high output uncertainty when dealing with format sensitive document, suggesting that reasoning over diverse reading pathways may improve OCR performance. To address this, we propose format decoupled reinforcement learning (FD-RL), which leverages high-entropy patterns for targeted optimization. Our approach employs entropy-based data filtration strategy to identify format-intensive instances, and adopt format decoupled rewards tailored to different format types, enabling format-level validation rather than token-level memorization. FD-RL achieves an average score of 90.41 on OmniDocBench, setting a new record for end-to-end models on this highly popular benchmark. More importantly, we conduct comprehensive ablation studies over data, training, filtering, and rewarding strategies, thoroughly validating their effectiveness.

Chi Zhang, Haibo Qiu, Qiming Zhang et al.

Reinforcement Learning with Verifiable Rewards (RLVR) has significantly advanced the reasoning capabilities of Large Language Models (LLMs) and is now being applied to Vision-Language Models (VLMs). However, vanilla RLVR for VLMs verifies only the final textual output, critically neglecting the foundational step of visual perception. This oversight leads to visual hallucinations and reward hacking, as reasoning built upon flawed perception is inherently unreliable. To address this, we propose PEARL (Perceptual-Evidence Anchored Reinforced Learning), a dual-branch, perception-reasoning synergistic that strengthens multimodal reasoning by explicitly anchoring it to verified visual evidence. For each reasoning-oriented QA instance, PEARL first derive a perception checklist -- a set of perception-oriented sub-questions with verifiable answers that probe the model's understanding of key visual evidence. During training, auxiliary rollouts on this checklist yield a perceptual reward that both directly reinforces the model's perception ability and acts as a fidelity gate for reasoning. If the model passes the perception check, its policy update is biased towards evidence-anchored reasoning. Otherwise, the process is halted to prevent reasoning from flawed premises. PEARL can be seamlessly integrated with popular RL methods like GRPO and DAPO. Comprehensive experiments show PEARL achieves substantial gains on multimodal reasoning benchmarks, e.g., a +9.7% improvement over the baseline and +6.6% over GRPO on MathVerse.

Kun Chen, Peng Shi, Haibo Qiu et al.

Reinforcement learning (RL) with verifiable rewards has recently catalyzed a wave of "MLLM-r1" approaches that bring RL to vision language models. Most representative paradigms begin with a cold start, typically employing supervised fine-tuning (SFT), to initialize the policy before RL. However, SFT-based cold start adopts the reasoning paradigm intertwined with task solution and output format, which may induce instruction-style overfitting, weakens out-of-distribution generalization, and ultimately affects downstream RL. We revisit the cold start along two views, its training method and data construction, and introduce the Generalization Factor (GF) coefficient to quantify the generalization capability under different methods. Our empirical study finds that preference-based training methods (e.g. DPO) generalizes better than SFT-based methods in cold start. Motivated by this, we propose SPECS-a Self-distilled, Preference-based Cold Start framework that decouples multimodal learning: (1) generates introspective preference data pairs via self-distillation, avoiding reliance on larger teachers or manual annotation; (2) performs preference-based training to learn, focusing on shallow, transferable surface-form criteria (format, structure, style) rather than memorizing content; and (3) hands off to RL with verifiable rewards for deep reasoning results. Experimental results across multiple multimodal benchmarks show that our decoupling learning framework yields consistent performance gains over strong baselines, improving MEGA-Bench by 4.1% and MathVista by 12.2%. Additional experiments indicate that SPECS contributes to reducing in-distribution "stuckness," improving exploration, stabilizing training, and raising the performance ceiling.

Shuai Fu, Jian Zhou, Qi Chen et al.

Diffusion probabilistic models (DPMs) have demonstrated remarkable progress in generative tasks, such as image and video synthesis. However, they still often produce hallucinated samples (hallucinations) that conflict with real-world knowledge, such as generating an implausible duplicate cup floating beside another cup. Despite their prevalence, the lack of feasible methodologies for systematically quantifying such hallucinations hinders progress in addressing this challenge and obscures potential pathways for designing next-generation generative models under factual constraints. In this work, we bridge this gap by focusing on a specific form of hallucination, which we term counting hallucination, referring to the generation of an incorrect number of instances or structured objects, such as a hand image with six fingers, despite such patterns being absent from the training data. To this end, we construct a dataset suite CountHalluSet, with well-defined counting criteria, comprising ToyShape, SimObject, and RealHand. Using these datasets, we develop a standardized evaluation protocol for quantifying counting hallucinations, and systematically examine how different sampling conditions in DPMs, including solver type, ODE solver order, sampling steps, and initial noise, affect counting hallucination levels. Furthermore, we analyze their correlation with common evaluation metrics such as FID, revealing that this widely used image quality metric fails to capture counting hallucinations consistently. This work aims to take the first step toward systematically quantifying hallucinations in diffusion models and offer new insights into the investigation of hallucination phenomena in image generation.

Xiaoxiao Ma, Feng Zhao, Pengyang Ling et al.

In this work, we first revisit the sampling issues in current autoregressive (AR) image generation models and identify that image tokens, unlike text tokens, exhibit lower information density and non-uniform spatial distribution. Accordingly, we present an entropy-informed decoding strategy that facilitates higher autoregressive generation quality with faster synthesis speed. Specifically, the proposed method introduces two main innovations: 1) dynamic temperature control guided by spatial entropy of token distributions, enhancing the balance between content diversity, alignment accuracy, and structural coherence in both mask-based and scale-wise models, without extra computational overhead, and 2) entropy-aware acceptance rules in speculative decoding, achieving near-lossless generation at about 85\% of the inference cost of conventional acceleration methods. Extensive experiments across multiple benchmarks using diverse AR image generation models demonstrate the effectiveness and generalizability of our approach in enhancing both generation quality and sampling speed.

Xiaoxiao Ma, Haibo Qiu, Guohui Zhang et al.

Reinforcement learning has recently been explored to improve text-to-image generation, yet applying existing GRPO algorithms to autoregressive (AR) image models remains challenging. The instability of the training process easily disrupts the pretrained model capability during long runs, resulting in marginal gains, degraded image quality, and poor generalization. In this work, we revisit GRPO for AR image generation and identify two key issues: contradictory gradients from unnecessary tokens and unstable policy entropy dynamics. To address these, we introduce STAGE, a stable and generalizable framework that leverages two targeted solutions: 1) Advantage/KL reweighting. Similarity-aware reweighting to alleviate conflicting updates; and 2) Entropy reward. An entropy-based reward corresponding to reference model to stabilize learning. With the help of alleviating conflicts between tokens and an entropy reward for stabilizing training, we reduce disruption of the pretrained distribution and mitigate reward hacking, which in turn improves generalization and transfer better to other benchmarks. Experiments across multiple benchmarks show that STAGE consistently improves visual quality, stability, and cross-task generalization compared to baseline GRPO.

Yufeng Zhong, Zhixiong Zeng, Lei Chen et al.

Optical Character Recognition (OCR) for mathematical formula is essential for the intelligent analysis of scientific literature. However, both task-specific and general vision-language models often struggle to handle the structural diversity, complexity, and real-world variability inherent in mathematical content. In this work, we present DocTron-Formula, a unified framework built upon general vision-language models, thereby eliminating the need for specialized architectures. Furthermore, we introduce CSFormula, a large-scale and challenging dataset that encompasses multidisciplinary and structurally complex formulas at the line, paragraph, and page levels. Through straightforward supervised fine-tuning, our approach achieves state-of-the-art performance across a variety of styles, scientific domains, and complex layouts. Experimental results demonstrate that our method not only surpasses specialized models in terms of accuracy and robustness, but also establishes a new paradigm for the automated understanding of complex scientific documents.

Zhixiong Zeng, Wenji Mao

Cross-Modal Retrieval (CMR) is an important research topic across multimodal computing and information retrieval, which takes one type of data as the query to retrieve relevant data of another type. It has been widely used in many real-world applications. Recently, the vision-language pre-trained models represented by CLIP demonstrate its superiority in learning the visual and textual representations and gain impressive performance on various vision and language related tasks. Although CLIP as well as the previous pre-trained models have shown great performance improvement in the unsupervised CMR, the performance and impact of these pre-trained models on the supervised CMR were rarely explored due to the lack of common representation for the multimodal class-level associations. In this paper, we take CLIP as the current representative vision-language pre-trained model to conduct a comprehensive empirical study. We evaluate its performance and impact on the supervised CMR, and attempt to answer several key research questions. To this end, we first propose a novel model CLIP4CMR (CLIP enhanced network for Cross-Modal Retrieval) that employs the pre-trained CLIP as backbone network to perform the supervised CMR. Then by means of the CLIP4CMR framework, we revisit the design of different learning objectives in current CMR methods to provide new insights on model design. Moreover, we investigate the most concerned aspects in applying CMR, including the robustness to modality imbalance and sensitivity to hyper-parameters, to provide new perspectives for practical applications. Through extensive experiments, we show that CLIP4CMR achieves the SOTA results with prominent improvements on the benchmark datasets, and can be used as a fundamental framework to empirically study the key research issues of the supervised CMR, with significant implications for model design and practical considerations.

Guohai Xu, Hehong Chen, Feng-Lin Li et al.

Live streaming is becoming an increasingly popular trend of sales in E-commerce. The core of live-streaming sales is to encourage customers to purchase in an online broadcasting room. To enable customers to better understand a product without jumping out, we propose AliMe MKG, a multi-modal knowledge graph that aims at providing a cognitive profile for products, through which customers are able to seek information about and understand a product. Based on the MKG, we build an online live assistant that highlights product search, product exhibition and question answering, allowing customers to skim over item list, view item details, and ask item-related questions. Our system has been launched online in the Taobao app, and currently serves hundreds of thousands of customers per day.