Can Ma

Shangpin Peng, Gengluo Li, Xingyu Wan et al.

Charts are a primary medium for conveying quantitative and relational information, yet systematically evaluating chart parsing models remains difficult. Existing benchmarks focus on narrow chart types and leave diagrammatic structures such as flowcharts and mind maps largely unaddressed, while models produce outputs in incompatible formats, and datasets rarely include the printed or hand-drawn images encountered in practice. To address these issues, we introduce ChartArena, a comprehensive bilingual benchmark covering eight chart families spanning both numeric charts and diagrammatic structures, each evaluated across three visual scenarios: digital renderings, printed photos, and hand-drawn photos. The dataset is built via a human-agent collaborative annotation pipeline with multi-stage human verification to ensure annotation reliability. To enable fair cross-model comparison, we further design a format-agnostic evaluation protocol that maps heterogeneous outputs into two canonical semantic spaces, a normalized triple view and a directed graph view, and scores them with structure-aware metrics. Through extensive evaluation of 26 leading MLLMs, we observe three consistent findings: (i) frontier proprietary models such as Gemini 3.1 Pro lead overall, yet the strongest open-source systems are rapidly closing the gap; (ii) document parsing models handle numeric charts reasonably but fall sharply behind on diagrammatic structures; and (iii) expert chart parsers remain limited to narrow chart families. Across all models, radar charts and hand-drawn scenarios stay especially challenging. These findings show that ChartArena exposes clear capability gaps and provides a unified foundation for future progress. ChartArena is publicly available at https://github.com/pspdada/ChartArena.

Jiahao Lyu, Pei Fu, Zhenhang Li et al.

End-to-end In-Image Machine Translation (IIMT) aims to convert text embedded within an image into a target language while preserving the original visual context, layout, and rendering style. However, existing IIMT benchmarks are largely synthetic and thus fail to reflect real-world complexity, while current evaluation protocols focus on single-modality metrics and overlook cross-modal faithfulness between rendered text and model outputs. To address these shortcomings, we present In-image Machine Translation Benchmark (IMTBench), a new benchmark of 2,500 image translation samples covering four practical scenarios and nine languages. IMTBench supports multi-aspect evaluation, including translation quality, background preservation, overall image quality, and a cross-modal alignment score that measures consistency between the translated text produced by the model and the text rendered in the translated image. We benchmark strong commercial cascade systems, and both closed- and open-source unified multi-modal models, and observe large performance gaps across scenarios and languages, especially on natural scenes and resource-limited languages, highlighting substantial headroom for end-to-end image text translation. We hope IMTBench establishes a standardized benchmark to accelerate progress in this emerging task.

Shihao Rao, Liang Li, Jiapeng Liu et al.

Recent advances in large language models (LLMs) have opened up new possibilities for automated document formatting. However, real-world formatting often requires identifying targets based on document content. This content-aware setting remains challenging and underexplored, primarily due to the lack of dedicated evaluation datasets.To enable evaluation in realistic content-aware scenarios, we introduce DocFormBench, a benchmark that extends Text-to-Format evaluation to diverse formatting requirements, along with metrics for both accuracy and efficiency.To mitigate redundant document reading in existing methods during formatting, we propose DocFormFlow, a workflow formatting method that decouples target localization from modification execution into what to format and how. Extensive experiments across multiple LLMs and multimodal models show that DocFormFlow consistently improves formatting accuracy while reducing token consumption compared to representative baselines. Further analysis reveals that precise target localization is the primary factor influencing formatting performance. We hope DocFormBench and DocFormFlow will facilitate future research toward more intelligent and reliable document formatting.

Daiqing Wu, Xuan Zhang, Dongbao Yang et al.

The maturation of Large Audio Language Models (LALMs) has raised growing expectations for them to comprehend complex audio much like humans. Current efforts primarily replicate text-based reasoning by contextualizing audio content through a one-time encoding, which introduces a critical information bottleneck. Drawing inspiration from human cognition, we propose audio-interleaved reasoning to break through this bottleneck. It treats audio as an active reasoning component, enabling sustained audio engagement and perception-grounded analysis. To instantiate it, we introduce a two-stage training framework, first teaching LALMs to localize salient audio segments through supervised fine-tuning, and then incentivizing proficient re-listening via reinforcement learning. In parallel, a structured data generation pipeline is developed to produce high-quality training data. Consequently, we present Echo, a LALM capable of dynamically re-listening to audio in demand during reasoning. On audio comprehension benchmarks, Echo achieves overall superiority in both challenging expert-level and general-purpose tasks. Comprehensive analysis further confirms the efficiency and generalizability of audio-interleaved reasoning, establishing it as a promising direction for advancing audio comprehension. Project page: https://github.com/wdqqdw/Echo.

Xunyu Zhu, Jian Li, Yong Liu et al.

Large Language Models (LLMs) have transformed natural language processing tasks successfully. Yet, their large size and high computational needs pose challenges for practical use, especially in resource-limited settings. Model compression has emerged as a key research area to address these challenges. This paper presents a survey of model compression techniques for LLMs. We cover methods like quantization, pruning, and knowledge distillation, highlighting recent advancements. We also discuss benchmarking strategies and evaluation metrics crucial for assessing compressed LLMs. This survey offers valuable insights for researchers and practitioners, aiming to enhance efficiency and real-world applicability of LLMs while laying a foundation for future advancements.

Chengyang Fang, Jiangnan Li, Liang Li et al.

Text-based Visual Question Answering (TextVQA) aims at answering questions about the text in images. Most works in this field focus on designing network structures or pre-training tasks. All these methods list the OCR texts in reading order (from left to right and top to bottom) to form a sequence, which is treated as a natural language ``sentence''. However, they ignore the fact that most OCR words in the TextVQA task do not have a semantical contextual relationship. In addition, these approaches use 1-D position embedding to construct the spatial relation between OCR tokens sequentially, which is not reasonable. The 1-D position embedding can only represent the left-right sequence relationship between words in a sentence, but not the complex spatial position relationship. To tackle these problems, we propose a novel method named Separate and Locate (SaL) that explores text contextual cues and designs spatial position embedding to construct spatial relations between OCR texts. Specifically, we propose a Text Semantic Separate (TSS) module that helps the model recognize whether words have semantic contextual relations. Then, we introduce a Spatial Circle Position (SCP) module that helps the model better construct and reason the spatial position relationships between OCR texts. Our SaL model outperforms the baseline model by 4.44% and 3.96% accuracy on TextVQA and ST-VQA datasets. Compared with the pre-training state-of-the-art method pre-trained on 64 million pre-training samples, our method, without any pre-training tasks, still achieves 2.68% and 2.52% accuracy improvement on TextVQA and ST-VQA. Our code and models will be released at https://github.com/fangbufang/SaL.

Liang Li, Ruiying Geng, Chengyang Fang et al.

There are three problems existing in the popular data-to-text datasets. First, the large-scale datasets either contain noise or lack real application scenarios. Second, the datasets close to real applications are relatively small in size. Last, current datasets bias in the English language while leaving other languages underexplored. To alleviate these limitations, in this paper, we present CATS, a pragmatic Chinese answer-to-sequence dataset with large scale and high quality. The dataset aims to generate textual descriptions for the answer in the practical TableQA system. Further, to bridge the structural gap between the input SQL and table and establish better semantic alignments, we propose a Unified Graph Transformation approach to establish a joint encoding space for the two hybrid knowledge resources and convert this task to a graph-to-text problem. The experiment results demonstrate the effectiveness of our proposed method. Further analysis on CATS attests to both the high quality and challenges of the dataset.

Gengluo Li, Shangpin Peng, Xingyu Wan et al.

Vision Large Language Models (VLLMs) have achieved remarkable success in modern text-rich visual understanding. However, their perceptual robustness in the face of the continuous morphological evolution of historical writing systems remains largely unexplored. Existing ancient text datasets typically focus on isolated historical periods, failing to capture the systematic visual distribution shifts spanning thousands of years. To bridge this gap and empower Digital Humanities, we introduce Chronicles-OCR, the first comprehensive benchmark specifically designed to evaluate the cross-temporal visual perception capabilities of VLLMs across the complete evolutionary trajectory of Chinese characters, known as the Seven Chinese Scripts. Curated in collaboration with top-tier institutional domain experts, the dataset comprises 2,800 strictly balanced images encompassing highly diverse physical media, ranging from tortoise shells to paper-based calligraphy. To accommodate the drastic morphological and topological variations across different historical stages, we propose a novel Stage-Adaptive Annotation Paradigm. Based on this, Chronicles-OCR formulates four rigorous quantitative tasks: cross-period character spotting, fine-grained archaic character recognition via visual referring, ancient text parsing, and script classification. By isolating visual perception from semantic reasoning, Chronicles-OCR provides an authoritative platform to expose the limitations of current VLLMs, paving the way for robust, evolution-aware historical text perception. Chronicles-OCR is publicly available at https://github.com/VirtualLUOUCAS/Chronicles-OCR.

Chengyang Fang, Gangyan Zeng, Yu Zhou et al.

Texts in scene images convey critical information for scene understanding and reasoning. The abilities of reading and reasoning matter for the model in the text-based visual question answering (TextVQA) process. However, current TextVQA models do not center on the text and suffer from several limitations. The model is easily dominated by language biases and optical character recognition (OCR) errors due to the absence of semantic guidance in the answer prediction process. In this paper, we propose a novel Semantics-Centered Network (SC-Net) that consists of an instance-level contrastive semantic prediction module (ICSP) and a semantics-centered transformer module (SCT). Equipped with the two modules, the semantics-centered model can resist the language biases and the accumulated errors from OCR. Extensive experiments on TextVQA and ST-VQA datasets show the effectiveness of our model. SC-Net surpasses previous works with a noticeable margin and is more reasonable for the TextVQA task.

Daiqing Wu, Dongbao Yang, Huawen Shen et al.

With the proliferation of social media posts in recent years, the need to detect sentiments in multimodal (image-text) content has grown rapidly. Since posts are user-generated, the image and text from the same post can express different or even contradictory sentiments, leading to potential \textbf{sentiment discrepancy}. However, existing works mainly adopt a single-branch fusion structure that primarily captures the consistent sentiment between image and text. The ignorance or implicit modeling of discrepant sentiment results in compromised unimodal encoding and limited performance. In this paper, we propose a semantics Completion and Decomposition (CoDe) network to resolve the above issue. In the semantics completion module, we complement image and text representations with the semantics of the in-image text, helping bridge the sentiment gap. In the semantics decomposition module, we decompose image and text representations with exclusive projection and contrastive learning, thereby explicitly capturing the discrepant sentiment between modalities. Finally, we fuse image and text representations by cross-attention and combine them with the learned discrepant sentiment for final classification. Extensive experiments on four datasets demonstrate the superiority of CoDe and the effectiveness of each proposed module.

Gengluo Li, Chengquan Zhang, Yupu Liang et al.

Document parsing has recently advanced with multimodal large language models (MLLMs) that directly map document images to structured outputs. Traditional cascaded pipelines depend on precise layout analysis and often fail under casually captured or non-standard conditions. Although end-to-end approaches mitigate this dependency, they still exhibit repetitive, hallucinated, and structurally inconsistent predictions - primarily due to the scarcity of large-scale, high-quality full-page (document-level) end-to-end parsing data and the lack of structure-aware training strategies. To address these challenges, we propose a data-training co-design framework for robust end-to-end document parsing. A Realistic Scene Synthesis strategy constructs large-scale, structurally diverse full-page end-to-end supervision by composing layout templates with rich document elements, while a Document-Aware Training Recipe introduces progressive learning and structure-token optimization to enhance structural fidelity and decoding stability. We further build Wild-OmniDocBench, a benchmark derived from real-world captured documents for robustness evaluation. Integrated into a 1B-parameter MLLM, our method achieves superior accuracy and robustness across both scanned/digital and real-world captured scenarios. All models, data synthesis pipelines, and benchmarks will be publicly released to advance future research in document understanding.

Liang Li, Ruiying Geng, Chengyang Fang et al.

Table-to-text generation aims at automatically generating text to help people conveniently obtain salient information in tables. Recent works explicitly decompose the generation process into content planning and surface generation stages, employing two autoregressive networks for them respectively. However, they are computationally expensive due to the non-parallelizable nature of autoregressive decoding and the redundant parameters of two networks. In this paper, we propose the first totally non-autoregressive table-to-text model (Plan-then-Seam, PTS) that produces its outputs in parallel with one single network. PTS firstly writes and calibrates one plan of the content to be generated with a novel rethinking pointer predictor, and then takes the plan as the context for seaming to decode the description. These two steps share parameters and perform iteratively to capture token inter-dependency while keeping parallel decoding. Experiments on two public benchmarks show that PTS achieves 3.0~5.6 times speedup for inference time, reducing 50% parameters, while maintaining as least comparable performance against strong two-stage table-to-text competitors.

Liang Li, Ruiying Geng, Bowen Li et al.

Most graph-to-text works are built on the encoder-decoder framework with cross-attention mechanism. Recent studies have shown that explicitly modeling the input graph structure can significantly improve the performance. However, the vanilla structural encoder cannot capture all specialized information in a single forward pass for all decoding steps, resulting in inaccurate semantic representations. Meanwhile, the input graph is flatted as an unordered sequence in the cross attention, ignoring the original graph structure. As a result, the obtained input graph context vector in the decoder may be flawed. To address these issues, we propose a Structure-Aware Cross-Attention (SACA) mechanism to re-encode the input graph representation conditioning on the newly generated context at each decoding step in a structure aware manner. We further adapt SACA and introduce its variant Dynamic Graph Pruning (DGP) mechanism to dynamically drop irrelevant nodes in the decoding process. We achieve new state-of-the-art results on two graph-to-text datasets, LDC2020T02 and ENT-DESC, with only minor increase on computational cost.

Gengluo Li, Chengquan Zhang, Yupu Liang et al.

End-to-end text-image machine translation (TIMT), which directly translates textual content in images across languages, is crucial for real-world multilingual scene understanding. Despite advances in vision-language large models (VLLMs), robustness across diverse visual scenes and low-resource languages remains underexplored due to limited evaluation resources. We present MMTIT-Bench, a human-verified multilingual and multi-scenario benchmark with 1,400 images spanning fourteen non-English and non-Chinese languages and diverse settings such as documents, scenes, and web images, enabling rigorous assessment of end-to-end TIMT. Beyond benchmarking, we study how reasoning-oriented data design improves translation. Although recent VLLMs have begun to incorporate long Chain-of-Thought (CoT) reasoning, effective thinking paradigms for TIMT are still immature: existing designs either cascade parsing and translation in a sequential manner or focus on language-only reasoning, overlooking the visual cognition central to VLLMs. We propose Cognition-Perception-Reasoning for Translation (CPR-Trans), a data paradigm that integrates scene cognition, text perception, and translation reasoning within a unified reasoning process. Using a VLLM-driven data generation pipeline, CPR-Trans provides structured, interpretable supervision that aligns perception with reasoning. Experiments on 3B and 7B models show consistent gains in accuracy and interpretability. We will release MMTIT-Bench to promote the multilingual and multi-scenario TIMT research upon acceptance.

Shujie Li, Liang Li, Ruiying Geng et al.

Data-to-text (D2T) generation aims to transform structured data into natural language text. Data-to-text pre-training has proved to be powerful in enhancing D2T generation and yields impressive performances. However, previous pre-training methods either oversimplified structured data into a sequence without considering input structures or designed training objectives tailored for a specific data structure (e.g., table or knowledge graph). In this paper, we unify different types of structured data (i.e., table, key-value data, knowledge graph) into the graph format and cast different data-to-text generation tasks as graph-to-text generation. To effectively exploit the structural information of the input graph, we propose a structure-enhanced pre-training method for D2T generation by designing a structure-enhanced Transformer. Concretely, we devise a position matrix for the Transformer, encoding relative positional information of connected nodes in the input graph. In addition, we propose a new attention matrix to incorporate graph structures into the original Transformer by taking the available explicit connectivity structure into account. Extensive experiments on six benchmark datasets show the effectiveness of our model. Our source codes are available at https://github.com/AlibabaResearch/DAMO-ConvAI/tree/main/unid2t.

Huawen Shen, Chang Liu, Gengluo Li et al.

Pursuing human-like interaction for Graphical User Interface (GUI) agents requires understanding the GUI context and following user instructions. However, existing works typically couple these two aspects and focus more on instruct-following abilities, while ignoring the importance of understanding the GUI context. In this paper, we introduce an instruction-free GUI navigation dataset, termed Insight-UI Dataset, to enhance model comprehension of GUI environments. Insight-UI Dataset is automatically generated from the Common Crawl corpus, simulating various platforms -- including iOS, Android, Windows, and Linux -- across multiple resolutions on 312K domains. Although GUI interactions vary by context, diverse interfaces share common internal patterns, such as clicking an item to view its details. It implies the feasibility of independent GUI operation learning, followed by joint optimization with instruction tuning. Thereby, we develop the GUI agent model Falcon-UI, which is initially pretrained on Insight-UI Dataset and subsequently fine-tuned on Android and Web GUI datasets, including AITW, AITZ, Android Control, and Mind2Web. With 7 billion parameters, Falcon-UI achieves accuracy comparable to the 72 billion-parameter Qwen2VL on AITZ, validating the alignment between GUI context comprehension and agent performance. Our code and dataset will be open-sourced.

Xunyu Zhu, Jian Li, Can Ma et al.

Large Language Models (LLMs) have demonstrated exceptional proficiency in mathematical reasoning tasks due to their extensive parameter counts and training on vast datasets. Despite these capabilities, deploying LLMs is hindered by their computational demands. Distilling LLM mathematical reasoning into Smaller Language Models (SLMs) has emerged as a solution to this challenge, although these smaller models often suffer from errors in calculation and semantic understanding. Prior work has proposed Program-of-Thought Distillation (PoTD) to avoid calculation error. To further address semantic understanding errors, we propose Key-Point-Driven Mathematical Reasoning Distillation (KPDD). KPDD enhances the reasoning performance of SLMs by breaking down the problem-solving process into three stages: Core Question Extraction, Problem-Solving Information Extraction, and Step-by-Step Solution. This method is further divided into KPDD-CoT, which generates Chain-of-Thought rationales, and KPDD-PoT, which creates Program-of-Thought rationales. The experiment results show that KPDD-CoT significantly improves reasoning abilities, while KPDD-PoT achieves state-of-the-art performance in mathematical reasoning tasks. Our approach effectively mitigates misunderstanding errors, advancing the deployment of efficient and capable SLMs.

Weixin Guan, Liang Li, Jiapeng Liu et al.

Large Reasoning Language Models (LRLMs) demonstrate impressive capabilities on complex tasks by utilizing long Chain-of-Thought reasoning. However, they are prone to overthinking, which generates redundant reasoning steps that degrade both performance and efficiency. Recently, early-exit strategies are proposed to mitigate overthinking by dynamically and adaptively terminating redundant reasoning. However, current early-exit methods either introduce extra training overhead by relying on proxy models or limit inference throughput due to the frequent content switching between reasoning and generating probing answers. Moreover, most early-exit methods harm LRLMs performance due to over-truncation. Our insight stems from an observation: overthinking often causes LRLMs to deviate from the correct reasoning path, which is frequently accompanied by high-entropy transition tokens. Given this, we propose an early-exit method deeply coupled with the native reasoning process, which leverages the path deviation index as a dedicated monitoring metric for the frequent occurrence of high-entropy transition tokens to dynamically detect and terminate overthinking trajectories. We conduct experiments across multiple benchmarks using LRLMs of different types and scales, and the results indicate that our method delivers the largest performance improvement over vanilla CoT compared to existing early-exit methods.

Daiqing Wu, Dongbao Yang, Yu Zhou et al.

Visual emotion recognition (VER) is a longstanding field that has garnered increasing attention with the advancement of deep neural networks. Although recent studies have achieved notable improvements by leveraging the knowledge embedded within pre-trained visual models, the lack of direct association between factual-level features and emotional categories, called the "affective gap", limits the applicability of pre-training knowledge for VER tasks. On the contrary, the explicit emotional expression and high information density in textual modality eliminate the "affective gap". Therefore, we propose borrowing the knowledge from the pre-trained textual model to enhance the emotional perception of pre-trained visual models. We focus on the factual and emotional connections between images and texts in noisy social media data, and propose Partitioned Adaptive Contrastive Learning (PACL) to leverage these connections. Specifically, we manage to separate different types of samples and devise distinct contrastive learning strategies for each type. By dynamically constructing negative and positive pairs, we fully exploit the potential of noisy samples. Through comprehensive experiments, we demonstrate that bridging the "affective gap" significantly improves the performance of various pre-trained visual models in downstream emotion-related tasks. Our code is released on https://github.com/wdqqdw/PACL.

Daiqing Wu, Dongbao Yang, Sicheng Zhao et al.

Recently, Multimodal Large Language Models (MLLMs) have achieved exceptional performance across diverse tasks, continually surpassing previous expectations regarding their capabilities. Nevertheless, their proficiency in perceiving emotions from images remains debated, with studies yielding divergent results in zero-shot scenarios. We argue that this inconsistency stems partly from constraints in existing evaluation methods, including the oversight of plausible responses, limited emotional taxonomies, neglect of contextual factors, and labor-intensive annotations. To facilitate customized visual emotion evaluation for MLLMs, we propose an Emotion Statement Judgment task that overcomes these constraints. Complementing this task, we devise an automated pipeline that efficiently constructs emotion-centric statements with minimal human effort. Through systematically evaluating prevailing MLLMs, our study showcases their stronger performance in emotion interpretation and context-based emotion judgment, while revealing relative limitations in comprehending perception subjectivity. When compared to humans, even top-performing MLLMs like GPT4o demonstrate remarkable performance gaps, underscoring key areas for future improvement. By developing a fundamental evaluation framework and conducting a comprehensive MLLM assessment, we hope this work contributes to advancing emotional intelligence in MLLMs. Project page: https://github.com/wdqqdw/MVEI.

Yan Zhang, Gangyan Zeng, Daiqing Wu et al.

Video text-based visual question answering (Video TextVQA) aims to answer questions by explicitly reading and reasoning about the text involved in a video. Most works in this field follow a frame-level framework which suffers from redundant text entities and implicit relation modeling, resulting in limitations in both accuracy and efficiency. In this paper, we rethink the Video TextVQA task from an instance-oriented perspective and propose a novel model termed GAT (Gather and Trace). First, to obtain accurate reading result for each video text instance, a context-aggregated instance gathering module is designed to integrate the visual appearance, layout characteristics, and textual contents of the related entities into a unified textual representation. Then, to capture dynamic evolution of text in the video flow, an instance-focused trajectory tracing module is utilized to establish spatio-temporal relationships between instances and infer the final answer. Extensive experiments on several public Video TextVQA datasets validate the effectiveness and generalization of our framework. GAT outperforms existing Video TextVQA methods, video-language pretraining methods, and video large language models in both accuracy and inference speed. Notably, GAT surpasses the previous state-of-the-art Video TextVQA methods by 3.86\% in accuracy and achieves ten times of faster inference speed than video large language models. The source code is available at https://github.com/zhangyan-ucas/GAT.

Jiapeng Liu, Liang Li, Bing Li et al.

Recent works focus on synthesizing Chart Understanding (ChartU) training sets to inject advanced chart knowledge into Multimodal Large Language Models (MLLMs), where the sufficiency of the knowledge is typically verified by quantifying capability gains via the fine-tune-then-evaluate paradigm. However, full-set fine-tuning MLLMs to assess such gains incurs significant time costs, hindering the iterative refinement cycles of the ChartU dataset. Reviewing the ChartU dataset synthesis and data selection domains, we find that subsets can potentially probe the MLLMs' capability gains from full-set fine-tuning. Given that data diversity is vital for boosting MLLMs' performance and entropy reflects this feature, we propose EXaMCaP, which uses entropy gain maximization to select a subset. To obtain a high-diversity subset, EXaMCaP chooses the maximum-entropy subset from the large ChartU dataset. As enumerating all possible subsets is impractical, EXaMCaP iteratively selects samples to maximize the gain in set entropy relative to the current set, approximating the maximum-entropy subset of the full dataset. Experiments show that EXaMCaP outperforms baselines in probing the capability gains of the ChartU training set, along with its strong effectiveness across diverse subset sizes and compatibility with various MLLM architectures.

Yan Zhang, Daiqing Wu, Huawen Shen et al.

Graphical User Interface (GUI) grounding maps natural language instructions to the visual coordinates of target elements and serves as a core capability for autonomous GUI agents. Recent reinforcement learning methods (e.g., GRPO) have achieved strong performance, but they rely on expensive multiple rollouts and suffer from sparse signals on hard samples. These limitations make on-policy self-distillation (OPSD), which provides dense token-level supervision from a single rollout, a promising alternative. However, its applicability to GUI grounding remains unexplored. In this paper, we present GUI-SD, the first OPSD framework tailored for GUI grounding. First, it constructs a visually enriched privileged context for the teacher using a target bounding box and a Gaussian soft mask, providing informative guidance without leaking exact coordinates. Second, it employs entropy-guided distillation, which adaptively weights tokens based on digit significance and teacher confidence, concentrating optimization on the most impactful and reliable positions. Extensive experiments on six representative GUI grounding benchmarks show that GUI-SD consistently outperforms GRPO-based methods and naive OPSD in both accuracy and training efficiency. Code and training data are available at https://zhangyan-ucas.github.io/GUI-SD/.

Xunyu Zhu, Jian Li, Yong Liu et al.

This work addresses the challenge of democratizing advanced Large Language Models (LLMs) by compressing their mathematical reasoning capabilities into sub-billion parameter Small Language Models (SLMs) without compromising performance. We introduce Equation-of-Thought Distillation (EoTD), a novel technique that encapsulates the reasoning process into equation-based representations to construct an EoTD dataset for fine-tuning SLMs. Additionally, we propose the Ensemble Thoughts Distillation (ETD) framework to enhance the reasoning performance of SLMs. This involves creating a reasoning dataset with multiple thought processes, including Chain-of-Thought (CoT), Program-of-Thought (PoT), and Equation-of-Thought (EoT), and using it for fine-tuning. Our experimental performance demonstrates that EoTD significantly boosts the reasoning abilities of SLMs, while ETD enables these models to achieve state-of-the-art reasoning performance.

Yifei Zhang, Chang Liu, Jin Wei et al.

Text images are unique in their dual nature, encompassing both visual and linguistic information. The visual component encompasses structural and appearance-based features, while the linguistic dimension incorporates contextual and semantic elements. In scenarios with degraded visual quality, linguistic patterns serve as crucial supplements for comprehension, highlighting the necessity of integrating both aspects for robust scene text recognition (STR). Contemporary STR approaches often use language models or semantic reasoning modules to capture linguistic features, typically requiring large-scale annotated datasets. Self-supervised learning, which lacks annotations, presents challenges in disentangling linguistic features related to the global context. Typically, sequence contrastive learning emphasizes the alignment of local features, while masked image modeling (MIM) tends to exploit local structures to reconstruct visual patterns, resulting in limited linguistic knowledge. In this paper, we propose a Linguistics-aware Masked Image Modeling (LMIM) approach, which channels the linguistic information into the decoding process of MIM through a separate branch. Specifically, we design a linguistics alignment module to extract vision-independent features as linguistic guidance using inputs with different visual appearances. As features extend beyond mere visual structures, LMIM must consider the global context to achieve reconstruction. Extensive experiments on various benchmarks quantitatively demonstrate our state-of-the-art performance, and attention visualizations qualitatively show the simultaneous capture of both visual and linguistic information.

Yan Zhang, Gangyan Zeng, Huawen Shen et al.

Video text-based visual question answering (Video TextVQA) is a practical task that aims to answer questions by jointly reasoning textual and visual information in a given video. Inspired by the development of TextVQA in image domain, existing Video TextVQA approaches leverage a language model (e.g. T5) to process text-rich multiple frames and generate answers auto-regressively. Nevertheless, the spatio-temporal relationships among visual entities (including scene text and objects) will be disrupted and models are susceptible to interference from unrelated information, resulting in irrational reasoning and inaccurate answering. To tackle these challenges, we propose the TEA (stands for ``\textbf{T}rack th\textbf{E} \textbf{A}nswer'') method that better extends the generative TextVQA framework from image to video. TEA recovers the spatio-temporal relationships in a complementary way and incorporates OCR-aware clues to enhance the quality of reasoning questions. Extensive experiments on several public Video TextVQA datasets validate the effectiveness and generalization of our framework. TEA outperforms existing TextVQA methods, video-language pretraining methods and video large language models by great margins.

Daiqing Wu, Dongbao Yang, Sicheng Zhao et al.

The advancements in Multimodal Large Language Models (MLLMs) have enabled various multimodal tasks to be addressed under a zero-shot paradigm. This paradigm sidesteps the cost of model fine-tuning, emerging as a dominant trend in practical application. Nevertheless, Multimodal Sentiment Analysis (MSA), a pivotal challenge in the quest for general artificial intelligence, fails to accommodate this convenience. The zero-shot paradigm exhibits undesirable performance on MSA, casting doubt on whether MLLMs can perceive sentiments as competent as supervised models. By extending the zero-shot paradigm to In-Context Learning (ICL) and conducting an in-depth study on configuring demonstrations, we validate that MLLMs indeed possess such capability. Specifically, three key factors that cover demonstrations' retrieval, presentation, and distribution are comprehensively investigated and optimized. A sentimental predictive bias inherent in MLLMs is also discovered and later effectively counteracted. By complementing each other, the devised strategies for three factors result in average accuracy improvements of 15.9% on six MSA datasets against the zero-shot paradigm and 11.2% against the random ICL baseline.

Ruyue Liu, Rong Yin, Yong Liu et al.

Graph Contrastive Learning (GCL) has emerged as the foremost approach for self-supervised learning on graph-structured data. GCL reduces reliance on labeled data by learning robust representations from various augmented views. However, existing GCL methods typically depend on consistent stochastic augmentations, which overlook their impact on the intrinsic structure of the spectral domain, thereby limiting the model's ability to generalize effectively. To address these limitations, we propose a novel paradigm called AS-GCL that incorporates asymmetric spectral augmentation for graph contrastive learning. A typical GCL framework consists of three key components: graph data augmentation, view encoding, and contrastive loss. Our method introduces significant enhancements to each of these components. Specifically, for data augmentation, we apply spectral-based augmentation to minimize spectral variations, strengthen structural invariance, and reduce noise. With respect to encoding, we employ parameter-sharing encoders with distinct diffusion operators to generate diverse, noise-resistant graph views. For contrastive loss, we introduce an upper-bound loss function that promotes generalization by maintaining a balanced distribution of intra- and inter-class distance. To our knowledge, we are the first to encode augmentation views of the spectral domain using asymmetric encoders. Extensive experiments on eight benchmark datasets across various node-level tasks demonstrate the advantages of the proposed method.

Daiqing Wu, Dongbao Yang, Yu Zhou et al.

As posts on social media increase rapidly, analyzing the sentiments embedded in image-text pairs has become a popular research topic in recent years. Although existing works achieve impressive accomplishments in simultaneously harnessing image and text information, they lack the considerations of possible low-quality and missing modalities. In real-world applications, these issues might frequently occur, leading to urgent needs for models capable of predicting sentiment robustly. Therefore, we propose a Distribution-based feature Recovery and Fusion (DRF) method for robust multimodal sentiment analysis of image-text pairs. Specifically, we maintain a feature queue for each modality to approximate their feature distributions, through which we can simultaneously handle low-quality and missing modalities in a unified framework. For low-quality modalities, we reduce their contributions to the fusion by quantitatively estimating modality qualities based on the distributions. For missing modalities, we build inter-modal mapping relationships supervised by samples and distributions, thereby recovering the missing modalities from available ones. In experiments, two disruption strategies that corrupt and discard some modalities in samples are adopted to mimic the low-quality and missing modalities in various real-world scenarios. Through comprehensive experiments on three publicly available image-text datasets, we demonstrate the universal improvements of DRF compared to SOTA methods under both two strategies, validating its effectiveness in robust multimodal sentiment analysis.

Hanchang Cheng, Weimin Mu, Fan Liu et al.

Time series anomaly detection(TSAD) is a critical task in signal processing field, ensuring the reliability of complex systems. Reconstruction-based methods dominate in TSAD. Among these methods, VAE-based methods have achieved promising results. Existing VAE-based methods suffer from the limitation of single-window feature and insufficient leveraging of long-term time and frequency information. We propose a Conditional Variational AutoEncoder with Long-term dependency and Probabilistic time-frequency fusion, named LPCVAE. LPCVAE introduces LSTM to capture long-term dependencies beyond windows. It further incorporates a Product-of-Experts (PoE) mechanism for adaptive and distribution-level probabilistic fusion. This design effectively mitigates time-frequency information loss. Extensive experiments on four public datasets demonstrate it outperforms state-of-the-art methods. The results confirm that integrating long-term time and frequency representations with adaptive fusion yields a robust and efficient solution for TSAD.

Ruyue Liu, Rong Yin, Xiangzhen Bo et al.

Large scale pretrained models have revolutionized Natural Language Processing (NLP) and Computer Vision (CV), showcasing remarkable cross domain generalization abilities. However, in graph learning, models are typically trained on individual graph datasets, limiting their capacity to transfer knowledge across different graphs and tasks. This approach also heavily relies on large volumes of annotated data, which presents a significant challenge in resource-constrained settings. Unlike NLP and CV, graph structured data presents unique challenges due to its inherent heterogeneity, including domain specific feature spaces and structural diversity across various applications. To address these challenges, we propose a novel structure aware self supervised learning method for Text Attributed Graphs (SSTAG). By leveraging text as a unified representation medium for graph learning, SSTAG bridges the gap between the semantic reasoning of Large Language Models (LLMs) and the structural modeling capabilities of Graph Neural Networks (GNNs). Our approach introduces a dual knowledge distillation framework that co-distills both LLMs and GNNs into structure-aware multilayer perceptrons (MLPs), enhancing the scalability of large-scale TAGs. Additionally, we introduce an in-memory mechanism that stores typical graph representations, aligning them with memory anchors in an in-memory repository to integrate invariant knowledge, thereby improving the model's generalization ability. Extensive experiments demonstrate that SSTAG outperforms state-of-the-art models on cross-domain transfer learning tasks, achieves exceptional scalability, and reduces inference costs while maintaining competitive performance.

Rong Yin, Ruyue Liu, Xiaoshuai Hao et al.

Accurate extraction of molecular representations is a critical step in the drug discovery process. In recent years, significant progress has been made in molecular representation learning methods, among which multi-modal molecular representation methods based on images, and 2D/3D topologies have become increasingly mainstream. However, existing these multi-modal approaches often directly fuse information from different modalities, overlooking the potential of intermodal interactions and failing to adequately capture the complex higher-order relationships and invariant features between molecules. To overcome these challenges, we propose a structure-awareness-based multi-modal self-supervised molecular representation pre-training framework (MMSA) designed to enhance molecular graph representations by leveraging invariant knowledge between molecules. The framework consists of two main modules: the multi-modal molecular representation learning module and the structure-awareness module. The multi-modal molecular representation learning module collaboratively processes information from different modalities of the same molecule to overcome intermodal differences and generate a unified molecular embedding. Subsequently, the structure-awareness module enhances the molecular representation by constructing a hypergraph structure to model higher-order correlations between molecules. This module also introduces a memory mechanism for storing typical molecular representations, aligning them with memory anchors in the memory bank to integrate invariant knowledge, thereby improving the model generalization ability. Extensive experiments have demonstrated the effectiveness of MMSA, which achieves state-of-the-art performance on the MoleculeNet benchmark, with average ROC-AUC improvements ranging from 1.8% to 9.6% over baseline methods.

Ruyue Liu, Rong Yin, Xiangzhen Bo et al.

Federated graph learning (FGL) has gained significant attention for enabling heterogeneous clients to process their private graph data locally while interacting with a centralized server, thus maintaining privacy. However, graph data on clients are typically non-IID, posing a challenge for a single model to perform well across all clients. Another major bottleneck of FGL is the high cost of communication. To address these challenges, we propose a communication-efficient personalized federated graph learning algorithm, CEFGL. Our method decomposes the model parameters into low-rank generic and sparse private models. We employ a dual-channel encoder to learn sparse local knowledge in a personalized manner and low-rank global knowledge in a shared manner. Additionally, we perform multiple local stochastic gradient descent iterations between communication phases and integrate efficient compression techniques into the algorithm. The advantage of CEFGL lies in its ability to capture common and individual knowledge more precisely. By utilizing low-rank and sparse parameters along with compression techniques, CEFGL significantly reduces communication complexity. Extensive experiments demonstrate that our method achieves optimal classification accuracy in a variety of heterogeneous environments across sixteen datasets. Specifically, compared to the state-of-the-art method FedStar, the proposed method (with GIN as the base model) improves accuracy by 5.64\% on cross-datasets setting CHEM, reduces communication bits by a factor of 18.58, and reduces the communication time by a factor of 1.65.

Xunyu Zhu, Jian Li, Can Ma et al.

Large Language Models (LLMs) demonstrate exceptional reasoning capabilities, often achieving state-of-the-art performance in various tasks. However, their substantial computational and memory demands, due to billions of parameters, hinder deployment in resource-constrained environments. A promising solution is knowledge distillation, where LLMs transfer reasoning capabilities to Small Language Models (SLMs, $\le$ 1B parameters), enabling wider deployment on low-resource devices. Existing methods primarily focus on generating high-quality reasoning rationales for distillation datasets but often neglect the critical role of data quantity and quality. To address these challenges, we propose a Feedback-Driven Distillation (FDD) framework to enhance SLMs' mathematical reasoning capabilities. In the initialization stage, a distillation dataset is constructed by prompting LLMs to pair mathematical problems with corresponding reasoning rationales. We classify problems into easy and hard categories based on SLM performance. For easy problems, LLMs generate more complex variations, while for hard problems, new questions of similar complexity are synthesized. In addition, we propose a multi-round distillation paradigm to iteratively enrich the distillation datasets, thereby progressively improving the mathematical reasoning abilities of SLMs. Experimental results demonstrate that our method can make SLMs achieve SOTA mathematical reasoning performance.

Liang Li, Can Ma, Yinliang Yue et al.

Table-to-text generation aims at automatically generating natural text to help people to conveniently obtain the important information in tables. Although neural models for table-to-text have achieved remarkable progress, some problems still overlooked. The first is that the values recorded in many tables are mostly numbers in practice. The existing approaches do not do special treatment for these, and still regard these as words in natural language text. Secondly, the target texts in training dataset may contain redundant information or facts do not exist in the input tables. These may give wrong supervision signals to some methods based on content selection and planning and auxiliary supervision. To solve these problems, we propose two self-supervised tasks, Number Ordering and Significance Ordering, to help to learn better table representation. The former works on the column dimension to help to incorporate the size property of numbers into table representation. The latter acts on row dimension and help to learn a significance-aware table representation. We test our methods on the widely used dataset ROTOWIRE which consists of NBA game statistic and related news. The experimental results demonstrate that the model trained together with these two self-supervised tasks can generate text that contains more salient and well-organized facts, even without modeling context selection and planning. And we achieve the state-of-the-art performance on automatic metrics.

Dongbao Yang, Yu Zhou, Dayan Wu et al.

Modern object detection methods based on convolutional neural network suffer from severe catastrophic forgetting in learning new classes without original data. Due to time consumption, storage burden and privacy of old data, it is inadvisable to train the model from scratch with both old and new data when new object classes emerge after the model trained. In this paper, we propose a novel incremental object detector based on Faster R-CNN to continuously learn from new object classes without using old data. It is a triple network where an old model and a residual model as assistants for helping the incremental model learning on new classes without forgetting the previous learned knowledge. To better maintain the discrimination of features between old and new classes, the residual model is jointly trained on new classes in the incremental learning procedure. In addition, a corresponding distillation scheme is designed to guide the training process, which consists of a two-level residual distillation loss and a joint classification distillation loss. Extensive experiments on VOC2007 and COCO are conducted, and the results demonstrate that the proposed method can effectively learn to incrementally detect objects of new classes, and the problem of catastrophic forgetting is mitigated in this context.

Dezhao Luo, Chang Liu, Yu Zhou et al.

We propose a novel self-supervised method, referred to as Video Cloze Procedure (VCP), to learn rich spatial-temporal representations. VCP first generates "blanks" by withholding video clips and then creates "options" by applying spatio-temporal operations on the withheld clips. Finally, it fills the blanks with "options" and learns representations by predicting the categories of operations applied on the clips. VCP can act as either a proxy task or a target task in self-supervised learning. As a proxy task, it converts rich self-supervised representations into video clip operations (options), which enhances the flexibility and reduces the complexity of representation learning. As a target task, it can assess learned representation models in a uniform and interpretable manner. With VCP, we train spatial-temporal representation models (3D-CNNs) and apply such models on action recognition and video retrieval tasks. Experiments on commonly used benchmarks show that the trained models outperform the state-of-the-art self-supervised models with significant margins.