Zixuan Li

GLM-5 Team, Aohan Zeng, Xin Lv et al. · tsinghua

We present GLM-5, a next-generation foundation model designed to transition the paradigm of vibe coding to agentic engineering. Building upon the agentic, reasoning, and coding (ARC) capabilities of its predecessor, GLM-5 adopts DSA to significantly reduce training and inference costs while maintaining long-context fidelity. To advance model alignment and autonomy, we implement a new asynchronous reinforcement learning infrastructure that drastically improves post-training efficiency by decoupling generation from training. Furthermore, we propose novel asynchronous agent RL algorithms that further improve RL quality, enabling the model to learn from complex, long-horizon interactions more effectively. Through these innovations, GLM-5 achieves state-of-the-art performance on major open benchmarks. Most critically, GLM-5 demonstrates unprecedented capability in real-world coding tasks, surpassing previous baselines in handling end-to-end software engineering challenges. Code, models, and more information are available at https://github.com/zai-org/GLM-5.

Zixuan Li, Saiping Guan, Xiaolong Jin et al. · baidu

A Temporal Knowledge Graph (TKG) is a sequence of KGs corresponding to different timestamps. TKG reasoning aims to predict potential facts in the future given the historical KG sequences. One key of this task is to mine and understand evolutional patterns of facts from these sequences. The evolutional patterns are complex in two aspects, length-diversity and time-variability. Existing models for TKG reasoning focus on modeling fact sequences of a fixed length, which cannot discover complex evolutional patterns that vary in length. Furthermore, these models are all trained offline, which cannot well adapt to the changes of evolutional patterns from then on. Thus, we propose a new model, called Complex Evolutional Network (CEN), which uses a length-aware Convolutional Neural Network (CNN) to handle evolutional patterns of different lengths via an easy-to-difficult curriculum learning strategy. Besides, we propose to learn the model under the online setting so that it can adapt to the changes of evolutional patterns over time. Extensive experiments demonstrate that CEN obtains substantial performance improvement under both the traditional offline and the proposed online settings.

Weiwei Ding, Zixuan Li, Long Bai et al.

Knowledge Graphs (KGs) are widely used to mitigate the limitations of Large Language Models (LLMs), such as outdated knowledge and hallucinations. Existing LLM-KG integration frameworks typically rely on predefined operators to retrieve factual knowledge from KGs and inject it into prompts for answer generation. This paradigm faces two critical bottlenecks: 1) Inflexibility: The predefined operators are limited in scope and thus lack sufficient compositional expressiveness to fully capture the complex semantics required by KG questions. 2) Unscalability: Direct injection of factual knowledge into prompts limits scalability in handling large-scale factual knowledge. To address these two bottlenecks, we propose Code-on-Graph (CoG), a programmatic reasoning framework for LLM-KG integration. Specifically, given the factual knowledge retrieved at each reasoning step, CoG first identifies the corresponding KG schemas and represents these schemas as Python classes, which serve as abstract interfaces to the retrieved facts. It then generates executable code grounded in these classes, with the retrieved facts instantiated as objects of the corresponding classes during execution. This design enables flexible code-based reasoning while avoiding the direct injection of large-scale factual knowledge into prompts. Experiments on WebQSP, CWQ, and GrailQA demonstrate that CoG outperforms prior state-of-the-art models by up to 10.5%.

Zixuan Li, Zhongni Hou, Saiping Guan et al. · baidu

A Temporal Knowledge Graph (TKG) is a sequence of KGs with respective timestamps, which adopts quadruples in the form of (\emph{subject}, \emph{relation}, \emph{object}, \emph{timestamp}) to describe dynamic facts. TKG reasoning has facilitated many real-world applications via answering such queries as (\emph{query entity}, \emph{query relation}, \emph{?}, \emph{future timestamp}) about future. This is actually a matching task between a query and candidate entities based on their historical structures, which reflect behavioral trends of the entities at different timestamps. In addition, recent KGs provide background knowledge of all the entities, which is also helpful for the matching. Thus, in this paper, we propose the \textbf{Hi}storical \textbf{S}tructure \textbf{Match}ing (\textbf{HiSMatch}) model. It applies two structure encoders to capture the semantic information contained in the historical structures of the query and candidate entities. Besides, it adopts another encoder to integrate the background knowledge into the model. TKG reasoning experiments on six benchmark datasets demonstrate the significant improvement of the proposed HiSMatch model, with up to 5.6\% performance improvement in MRR, compared to the state-of-the-art baselines.

Long Chen, Yuchen Li, Chao Huang et al.

Interest in autonomous driving (AD) and intelligent vehicles (IVs) is growing at a rapid pace due to the convenience, safety, and economic benefits. Although a number of surveys have reviewed research achievements in this field, they are still limited in specific tasks, lack of systematic summary and research directions in the future. Here we propose a Survey of Surveys (SoS) for total technologies of AD and IVs that reviews the history, summarizes the milestones, and provides the perspectives, ethics, and future research directions. To our knowledge, this article is the first SoS with milestones in AD and IVs, which constitutes our complete research work together with two other technical surveys. We anticipate that this article will bring novel and diverse insights to researchers and abecedarians, and serve as a bridge between past and future.

Yucan Guo, Zixuan Li, Xiaolong Jin et al. · bytedance

Information Extraction (IE) aims to extract structural knowledge (e.g., entities, relations, events) from natural language texts, which brings challenges to existing methods due to task-specific schemas and complex text expressions. Code, as a typical kind of formalized language, is capable of describing structural knowledge under various schemas in a universal way. On the other hand, Large Language Models (LLMs) trained on both codes and texts have demonstrated powerful capabilities of transforming texts into codes, which provides a feasible solution to IE tasks. Therefore, in this paper, we propose a universal retrieval-augmented code generation framework based on LLMs, called Code4UIE, for IE tasks. Specifically, Code4UIE adopts Python classes to define task-specific schemas of various structural knowledge in a universal way. By so doing, extracting knowledge under these schemas can be transformed into generating codes that instantiate the predefined Python classes with the information in texts. To generate these codes more precisely, Code4UIE adopts the in-context learning mechanism to instruct LLMs with examples. In order to obtain appropriate examples for different tasks, Code4UIE explores several example retrieval strategies, which can retrieve examples semantically similar to the given texts. Extensive experiments on five representative IE tasks across nine datasets demonstrate the effectiveness of the Code4UIE framework.

Pedro Figueiredo, Zixuan Li, Beibei Wang et al.

We propose a neural formulation for estimating the appearance of complex luminaires. We focus on challenging luminaires with complex light transport (e.g., small emitters enclosed by multiple specular layers) that are difficult for (bidirectional) path tracing. To this end, we use light tracing to construct paths from emitters to the exit surfaces and formulate appearance estimation as a distribution learning problem. Specifically, we model the probability density function (pdf) of outgoing radiance on the exit surfaces using a large normalizing flow network, and recover the outgoing radiance as the product of the estimated pdf and flux. To enable efficient inference, we distill the learned appearance into a lightweight MLP that directly estimates radiance on the exit surfaces. We additionally train a sampling network for effective direct illumination computation from the luminaire, and a blending network to composite the luminaire into the scene. Our formulation makes it feasible to render challenging luminaires using low sample counts in arbitrary scenes.

Long Chen, Siyu Teng, Bai Li et al.

Growing interest in autonomous driving (AD) and intelligent vehicles (IVs) is fueled by their promise for enhanced safety, efficiency, and economic benefits. While previous surveys have captured progress in this field, a comprehensive and forward-looking summary is needed. Our work fills this gap through three distinct articles. The first part, a "Survey of Surveys" (SoS), outlines the history, surveys, ethics, and future directions of AD and IV technologies. The second part, "Milestones in Autonomous Driving and Intelligent Vehicles Part I: Control, Computing System Design, Communication, HD Map, Testing, and Human Behaviors" delves into the development of control, computing system, communication, HD map, testing, and human behaviors in IVs. This part, the third part, reviews perception and planning in the context of IVs. Aiming to provide a comprehensive overview of the latest advancements in AD and IVs, this work caters to both newcomers and seasoned researchers. By integrating the SoS and Part I, we offer unique insights and strive to serve as a bridge between past achievements and future possibilities in this dynamic field.

Weicheng Ren, Zixuan Li, Xiaolong Jin et al.

Nested Event Extraction (NEE) aims to extract complex event structures where an event contains other events as its arguments recursively. Nested events involve a kind of Pivot Elements (PEs) that simultaneously act as arguments of outer-nest events and as triggers of inner-nest events, and thus connect them into nested structures. This special characteristic of PEs brings challenges to existing NEE methods, as they cannot well cope with the dual identities of PEs. Therefore, this paper proposes a new model, called PerNee, which extracts nested events mainly based on recognizing PEs. Specifically, PerNee first recognizes the triggers of both inner-nest and outer-nest events and further recognizes the PEs via classifying the relation type between trigger pairs. The model uses prompt learning to incorporate information from both event types and argument roles for better trigger and argument representations to improve NEE performance. Since existing NEE datasets (e.g., Genia11) are limited to specific domains and contain a narrow range of event types with nested structures, we systematically categorize nested events in the generic domain and construct a new NEE dataset, called ACE2005-Nest. Experimental results demonstrate that PerNee consistently achieves state-of-the-art performance on ACE2005-Nest, Genia11, and Genia13. The ACE2005-Nest dataset and the code of the PerNee model are available at https://github.com/waysonren/PerNee.

Guhong Chen, Liyang Fan, Zihan Gong et al.

Current research in LLM-based simulation systems lacks comprehensive solutions for modeling real-world court proceedings, while existing legal language models struggle with dynamic courtroom interactions. We present AgentCourt, a comprehensive legal simulation framework that addresses these challenges through adversarial evolution of LLM-based agents. Our AgentCourt introduces a new adversarial evolutionary approach for agents called AdvEvol, which performs dynamic knowledge learning and evolution through structured adversarial interactions in a simulated courtroom program, breaking the limitations of the traditional reliance on static knowledge bases or manual annotations. By simulating 1,000 civil cases, we construct an evolving knowledge base that enhances the agents' legal reasoning abilities. The evolved lawyer agents demonstrated outstanding performance on our newly introduced CourtBench benchmark, achieving a 12.1% improvement in performance compared to the original lawyer agents. Evaluations by professional lawyers confirm the effectiveness of our approach across three critical dimensions: cognitive agility, professional knowledge, and logical rigor. Beyond outperforming specialized legal models in interactive reasoning tasks, our findings emphasize the importance of adversarial learning in legal AI and suggest promising directions for extending simulation-based legal reasoning to broader judicial and regulatory contexts. The project's code is available at: https://github.com/relic-yuexi/AgentCourt

Yuxuan Hu, Minghuan Tan, Chenwei Zhang et al.

Empathetic response generation is designed to comprehend the emotions of others and select the most appropriate strategies to assist them in resolving emotional challenges. Empathy can be categorized into cognitive empathy and affective empathy. The former pertains to the ability to understand and discern the emotional issues and situations of others, while the latter involves the capacity to provide comfort. To enhance one's empathetic abilities, it is essential to develop both these aspects. Therefore, we develop an innovative framework that combines retrieval augmentation and emotional support strategy integration. Our framework starts with the introduction of a comprehensive emotional palette for empathy. We then apply appraisal theory to decompose this palette and create a database of empathetic responses. This database serves as an external resource and enhances the LLM's empathy by integrating semantic retrieval mechanisms. Moreover, our framework places a strong emphasis on the proper articulation of response strategies. By incorporating emotional support strategies, we aim to enrich the model's capabilities in both cognitive and affective empathy, leading to a more nuanced and comprehensive empathetic response. Finally, we extract datasets ED and ET from the empathetic dialogue dataset \textsc{EmpatheticDialogues} and ExTES based on dialogue length. Experiments demonstrate that our framework can enhance the empathy ability of LLMs from both cognitive and affective empathy perspectives. Our code is released at https://github.com/CAS-SIAT-XinHai/APTNESS.

Long Bai, Saiping Guan, Zixuan Li et al.

Script is a kind of structured knowledge extracted from texts, which contains a sequence of events. Based on such knowledge, script event prediction aims to predict the subsequent event. To do so, two aspects should be considered for events, namely, event description (i.e., what the events should contain) and event encoding (i.e., how they should be encoded). Most existing methods describe an event by a verb together with only a few core arguments (i.e., subject, object, and indirect object), which are not precise. In addition, existing event encoders are limited to a fixed number of arguments, which are not flexible to deal with extra information. Thus, in this paper, we propose the Rich Event Prediction (REP) framework for script event prediction. Fundamentally, it is based on the proposed rich event description, which enriches the existing ones with three kinds of important information, namely, the senses of verbs, extra semantic roles, and types of participants. REP contains an event extractor to extract such information from texts. Based on the extracted rich information, a predictor then selects the most probable subsequent event. The core component of the predictor is a transformer-based event encoder to flexibly deal with an arbitrary number of arguments. Experimental results on the widely used Gigaword Corpus show the effectiveness of the proposed framework.

Xuhui Jiang, Chengjin Xu, Yinghan Shen et al.

The flourishing of knowledge graph applications has driven the need for entity alignment (EA) across KGs. However, the heterogeneity of practical KGs, characterized by differing scales, structures, and limited overlapping entities, greatly surpasses that of existing EA datasets. This discrepancy highlights an oversimplified heterogeneity in current EA datasets, which obstructs a full understanding of the advancements achieved by recent EA methods. In this paper, we study the performance of EA methods in practical settings, specifically focusing on the alignment of highly heterogeneous KGs (HHKGs). Firstly, we address the oversimplified heterogeneity settings of current datasets and propose two new HHKG datasets that closely mimic practical EA scenarios. Then, based on these datasets, we conduct extensive experiments to evaluate previous representative EA methods. Our findings reveal that, in aligning HHKGs, valuable structure information can hardly be exploited through message-passing and aggregation mechanisms. This phenomenon leads to inferior performance of existing EA methods, especially those based on GNNs. These findings shed light on the potential problems associated with the conventional application of GNN-based methods as a panacea for all EA datasets. Consequently, in light of these observations and to elucidate what EA methodology is genuinely beneficial in practical scenarios, we undertake an in-depth analysis by implementing a simple but effective approach: Simple-HHEA. This method adaptly integrates entity name, structure, and temporal information to navigate the challenges posed by HHKGs. Our experiment results conclude that the key to the future EA model design in practice lies in their adaptability and efficiency to varying information quality conditions, as well as their capability to capture patterns across HHKGs.

Zhilei Hu, Zixuan Li, Daozhu Xu et al.

Event Relation Extraction (ERE) aims to extract multiple kinds of relations among events in texts. However, existing methods singly categorize event relations as different classes, which are inadequately capturing the intrinsic semantics of these relations. To comprehensively understand their intrinsic semantics, in this paper, we obtain prototype representations for each type of event relation and propose a Prototype-Enhanced Matching (ProtoEM) framework for the joint extraction of multiple kinds of event relations. Specifically, ProtoEM extracts event relations in a two-step manner, i.e., prototype representing and prototype matching. In the first step, to capture the connotations of different event relations, ProtoEM utilizes examples to represent the prototypes corresponding to these relations. Subsequently, to capture the interdependence among event relations, it constructs a dependency graph for the prototypes corresponding to these relations and utilized a Graph Neural Network (GNN)-based module for modeling. In the second step, it obtains the representations of new event pairs and calculates their similarity with those prototypes obtained in the first step to evaluate which types of event relations they belong to. Experimental results on the MAVEN-ERE dataset demonstrate that the proposed ProtoEM framework can effectively represent the prototypes of event relations and further obtain a significant improvement over baseline models.

Zixuan Li, Chuanzhen Wang, Haotian Sun

Distributed AI systems face critical memory management challenges across computation, communication, and deployment layers. RRAM based in memory computing suffers from scalability limitations due to device non idealities and fixed array sizes. Decentralized AI frameworks struggle with memory efficiency across NAT constrained networks due to static routing that ignores computational load. Multi agent deployment systems tightly couple application logic with execution environments, preventing adaptive memory optimization. These challenges stem from a fundamental lack of coordinated memory management across architectural layers. We introduce Self Evolving Distributed Memory Architecture for Scalable AI Systems, a three layer framework that unifies memory management across computation, communication, and deployment. Our approach features (1) memory guided matrix processing with dynamic partitioning based on device characteristics, (2) memory aware peer selection considering network topology and computational capacity, and (3) runtime adaptive deployment optimization through continuous reconfiguration. The framework maintains dual memory systems tracking both long term performance patterns and short term workload statistics. Experiments on COCO 2017, ImageNet, and SQuAD show that our method achieves 87.3 percent memory utilization efficiency and 142.5 operations per second compared to Ray Distributed at 72.1 percent and 98.7 operations per second, while reducing communication latency by 30.2 percent to 171.2 milliseconds and improving resource utilization to 82.7 percent. Our contributions include coordinated memory management across three architectural layers, workload adaptive resource allocation, and a dual memory architecture enabling dynamic system optimization.

Pengzhen Ren, Kaidong Zhang, Hetao Zheng et al.

Considering how to make the model accurately understand and follow natural language instructions and perform actions consistent with world knowledge is a key challenge in robot manipulation. This mainly includes human fuzzy instruction reasoning and the following of physical knowledge. Therefore, the embodied intelligence agent must have the ability to model world knowledge from training data. However, most existing vision and language robot manipulation methods mainly operate in less realistic simulator and language settings and lack explicit modeling of world knowledge. To bridge this gap, we introduce a novel and simple robot manipulation framework, called Surfer. It is based on the world model, treats robot manipulation as a state transfer of the visual scene, and decouples it into two parts: action and scene. Then, the generalization ability of the model on new instructions and new scenes is enhanced by explicit modeling of the action and scene prediction in multi-modal information. In addition to the framework, we also built a robot manipulation simulator that supports full physics execution based on the MuJoCo physics engine. It can automatically generate demonstration training data and test data, effectively reducing labor costs. To conduct a comprehensive and systematic evaluation of the robot manipulation model in terms of language understanding and physical execution, we also created a robotic manipulation benchmark with progressive reasoning tasks, called SeaWave. It contains 4 levels of progressive reasoning tasks and can provide a standardized testing platform for embedded AI agents in multi-modal environments. On average, Surfer achieved a success rate of 54.74% on the defined four levels of manipulation tasks, exceeding the best baseline performance of 47.64%.

Yantao Liu, Zixuan Li, Xiaolong Jin et al.

The Knowledge Base Question Answering (KBQA) task aims to answer natural language questions based on a given knowledge base. Recently, Large Language Models (LLMs) have shown strong capabilities in language understanding and can be used to solve this task. In doing so, a major challenge for LLMs is to overcome the immensity and heterogeneity of knowledge base schemas.Existing methods bypass this challenge by initially employing LLMs to generate drafts of logic forms without schema-specific details.Then, an extra module is used to inject schema information to these drafts.In contrast, in this paper, we propose a simple In-Context Schema Understanding (ICSU) method that enables LLMs to directly understand schemas by leveraging in-context learning. Specifically, ICSU provides schema information to LLMs using schema-related annotated examples. We investigate three example retrieval strategies based on raw questions, anonymized questions, and generated SPARQL queries. Experimental results show that ICSU demonstrates competitive performance compared to baseline methods on both the KQA Pro and WebQSP datasets.

Miao Su, Yucan Guo, Zhongni Hou et al.

Memory enables Large Language Model (LLM) agents to perceive, store, and use information from past dialogues, which is essential for personalization. However, existing methods fail to properly model the temporal dimension of memory in two aspects: 1) Temporal inaccuracy: memories are organized by dialogue time rather than their actual occurrence time; 2) Temporal fragmentation: existing methods focus on point-wise memory, losing durative information that captures persistent states and evolving patterns. To address these limitations, we propose Temporal Semantic Memory (TSM), a memory framework that models semantic time for point-wise memory and supports the construction and utilization of durative memory. During memory construction, it first builds a semantic timeline rather than a dialogue one. Then, it consolidates temporally continuous and semantically related information into a durative memory. During memory utilization, it incorporates the query's temporal intent on the semantic timeline, enabling the retrieval of temporally appropriate durative memories and providing time-valid, duration-consistent context to support response generation. Experiments on LongMemEval and LoCoMo show that TSM consistently outperforms existing methods and achieves up to 12.2% absolute improvement in accuracy, demonstrating the effectiveness of the proposed method.

Li Kang, Heng Zhou, Xiufeng Song et al.

Recent advancements in multimodal large language models and vision-languageaction models have significantly driven progress in Embodied AI. As the field transitions toward more complex task scenarios, multi-agent system frameworks are becoming essential for achieving scalable, efficient, and collaborative solutions. This shift is fueled by three primary factors: increasing agent capabilities, enhancing system efficiency through task delegation, and enabling advanced human-agent interactions. To address the challenges posed by multi-agent collaboration, we propose the Multi-Agent Robotic System (MARS) Challenge, held at the NeurIPS 2025 Workshop on SpaVLE. The competition focuses on two critical areas: planning and control, where participants explore multi-agent embodied planning using vision-language models (VLMs) to coordinate tasks and policy execution to perform robotic manipulation in dynamic environments. By evaluating solutions submitted by participants, the challenge provides valuable insights into the design and coordination of embodied multi-agent systems, contributing to the future development of advanced collaborative AI systems.

Zixuan Li, Zixiong Wang, Jian Yang et al.

Traditional physically-based material models rely on analytically derived bidirectional reflectance distribution functions (BRDFs), typically by considering statistics of micro-primitives such as facets, flakes, or spheres, sometimes combined with multi-bounce interactions such as layering and multiple scattering. These derivations are often complex and model-specific. Once an analytic BRDF evaluation is defined, one still needs to design an importance sampling method for it and evaluate the probability density function (pdf) of that sampling distribution, requiring further model-specific derivations. We present PureSample: a novel neural BRDF representation that allows learning a material's appearance purely by sampling forward random walks on the microgeometry, which is usually straightforward to implement. Our representation allows for efficient BRDF evaluation, importance sampling, and pdf evaluation, for homogeneous as well as spatially varying materials. We achieve this by two learnable components: first, the sampling distribution is modeled using a flow matching neural network, which allows both importance sampling and pdf evaluation; second, we introduce a view-dependent albedo term, captured by a lightweight neural network, which allows for converting a pdf value to a BRDF value for any pair of view and light directions. We demonstrate PureSample on challenging materials, including various microgeometries, multi-layered materials, and multiple-scattering microfacet materials.

GLM-4. 5 Team, Aohan Zeng, Xin Lv et al.

We present GLM-4.5, an open-source Mixture-of-Experts (MoE) large language model with 355B total parameters and 32B activated parameters, featuring a hybrid reasoning method that supports both thinking and direct response modes. Through multi-stage training on 23T tokens and comprehensive post-training with expert model iteration and reinforcement learning, GLM-4.5 achieves strong performance across agentic, reasoning, and coding (ARC) tasks, scoring 70.1% on TAU-Bench, 91.0% on AIME 24, and 64.2% on SWE-bench Verified. With much fewer parameters than several competitors, GLM-4.5 ranks 3rd overall among all evaluated models and 2nd on agentic benchmarks. We release both GLM-4.5 (355B parameters) and a compact version, GLM-4.5-Air (106B parameters), to advance research in reasoning and agentic AI systems. Code, models, and more information are available at https://github.com/zai-org/GLM-4.5.

Tianxing Chen, Yuran Wang, Mingleyang Li et al.

Robotic manipulation policies have made rapid progress in recent years, yet most existing approaches give limited consideration to memory capabilities. Consequently, they struggle to solve tasks that require reasoning over historical observations and maintaining task-relevant information over time, which are common requirements in real-world manipulation scenarios. Although several memory-aware policies have been proposed, systematic evaluation of memory-dependent manipulation remains underexplored, and the relationship between architectural design choices and memory performance is still not well understood. To address this gap, we introduce RMBench, a simulation benchmark comprising 9 manipulation tasks that span multiple levels of memory complexity, enabling systematic evaluation of policy memory capabilities. We further propose Mem-0, a modular manipulation policy with explicit memory components designed to support controlled ablation studies. Through extensive simulation and real-world experiments, we identify memory-related limitations in existing policies and provide empirical insights into how architectural design choices influence memory performance. The website is available at https://rmbench.github.io/.

Tianxing Chen, Zanxin Chen, Baijun Chen et al.

Simulation-based data synthesis has emerged as a powerful paradigm for advancing real-world robotic manipulation. Yet existing datasets remain insufficient for robust bimanual manipulation due to (1) the lack of scalable task generation methods and (2) oversimplified simulation environments. We present RoboTwin 2.0, a scalable framework for automated, large-scale generation of diverse and realistic data, together with unified evaluation protocols for dual-arm manipulation. At its core is RoboTwin-OD, an object library of 731 instances across 147 categories with semantic and manipulation-relevant annotations. Building on this, we design an expert data synthesis pipeline that leverages multimodal language models (MLLMs) and simulation-in-the-loop refinement to automatically generate task-level execution code. To improve sim-to-real transfer, RoboTwin 2.0 applies structured domain randomization along five axes: clutter, lighting, background, tabletop height, and language, enhancing data diversity and policy robustness. The framework is instantiated across 50 dual-arm tasks and five robot embodiments. Empirically, it yields a 10.9% gain in code generation success rate. For downstream policy learning, a VLA model trained with synthetic data plus only 10 real demonstrations achieves a 367% relative improvement over the 10-demo baseline, while zero-shot models trained solely on synthetic data obtain a 228% gain. These results highlight the effectiveness of RoboTwin 2.0 in strengthening sim-to-real transfer and robustness to environmental variations. We release the data generator, benchmark, dataset, and code to support scalable research in robust bimanual manipulation. Project Page: https://robotwin-platform.github.io/, Code: https://github.com/robotwin-Platform/robotwin/.

Yuze Zheng, Zixuan Li, Xiangxian Li et al.

Image classification models often demonstrate unstable performance in real-world applications due to variations in image information, driven by differing visual perspectives of subject objects and lighting discrepancies. To mitigate these challenges, existing studies commonly incorporate additional modal information matching the visual data to regularize the model's learning process, enabling the extraction of high-quality visual features from complex image regions. Specifically, in the realm of multimodal learning, cross-modal alignment is recognized as an effective strategy, harmonizing different modal information by learning a domain-consistent latent feature space for visual and semantic features. However, this approach may face limitations due to the heterogeneity between multimodal information, such as differences in feature distribution and structure. To address this issue, we introduce a Multimodal Alignment and Reconstruction Network (MARNet), designed to enhance the model's resistance to visual noise. Importantly, MARNet includes a cross-modal diffusion reconstruction module for smoothly and stably blending information across different domains. Experiments conducted on two benchmark datasets, Vireo-Food172 and Ingredient-101, demonstrate that MARNet effectively improves the quality of image information extracted by the model. It is a plug-and-play framework that can be rapidly integrated into various image classification frameworks, boosting model performance.

Di Zhu, Zixuan Li

Distributional metrics such as Fréchet Audio Distance cannot score individual music clips and correlate poorly with human judgments, while the only per-sample learned metric achieving high human correlation is closed-source. We introduce MUQ-EVAL, an open-source per-sample quality metric for AIgenerated music built by training lightweight prediction heads on frozen MuQ-310M features using MusicEval, a dataset of generated clips from 31 text-to-music systems with expert quality ratings. Our simplest model, frozen features with attention pooling and a two-layer MLP, achieves system-level SRCC = 0.957 and utterance-level SRCC = 0.838 with human mean opinion scores. A systematic ablation over training objectives and adaptation strategies shows that no addition meaningfully improves the frozen baseline, indicating that frozen MuQ representations already capture quality-relevant information. Encoder choice is the dominant design factor, outweighing all architectural and training decisions. LoRA-adapted models trained on as few as 150 clips already achieve usable correlation, enabling personalized quality evaluators from individual listener annotations. A controlled degradation analysis reveals selective sensitivity to signal-level artifacts but insensitivity to musical-structural distortions. Our metric, MUQ-EVAL, is fully open-source, outperforms existing open per-sample metrics, and runs in real time on a single consumer GPU. Code, model weights, and evaluation scripts are available at https://github.com/dgtql/MuQ-Eval.

Kaixuan Wang, Tianxing Chen, Jiawei Liu et al.

Learning in simulation provides a useful foundation for scaling robotic manipulation capabilities. However, this paradigm often suffers from a lack of data-generation-ready digital assets, in both scale and diversity. In this work, we present ManiTwin, an automated and efficient pipeline for generating data-generation-ready digital object twins. Our pipeline transforms a single image into simulation-ready and semantically annotated 3D asset, enabling large-scale robotic manipulation data generation. Using this pipeline, we construct ManiTwin-100K, a dataset containing 100K high-quality annotated 3D assets. Each asset is equipped with physical properties, language descriptions, functional annotations, and verified manipulation proposals. Experiments demonstrate that ManiTwin provides an efficient asset synthesis and annotation workflow, and that ManiTwin-100K offers high-quality and diverse assets for manipulation data generation, random scene synthesis, and VQA data generation, establishing a strong foundation for scalable simulation data synthesis and policy learning. Our webpage is available at https://manitwin.github.io/.

Zixuan Li, Haiying Lin, Zhangyu Wang et al.

Ground segmentation, as the basic task of unmanned intelligent perception, provides an important support for the target detection task. Unstructured road scenes represented by open-pit mines have irregular boundary lines and uneven road surfaces, which lead to segmentation errors in current ground segmentation methods. To solve this problem, a ground segmentation method based on point cloud map is proposed, which involves three parts: region of interest extraction, point cloud registration and background subtraction. Firstly, establishing boundary semantic associations to obtain regions of interest in unstructured roads. Secondly, establishing the location association between point cloud map and the real-time point cloud of region of interest by semantics information. Thirdly, establishing a background model based on Gaussian distribution according to location association, and segments the ground in real-time point cloud by the background substraction method. Experimental results show that the correct segmentation rate of ground points is 99.95%, and the running time is 26ms. Compared with state of the art ground segmentation algorithm Patchwork++, the average accuracy of ground point segmentation is increased by 7.43%, and the running time is increased by 17ms. Furthermore, the proposed method is practically applied to unstructured road scenarios represented by open pit mines.

Wenxuan Liu, Zixuan Li, Bai Long et al.

Deep Research (DR) requires LLM agents to autonomously perform multi-step information seeking, processing, and reasoning to generate comprehensive reports. In contrast to existing studies that mainly focus on unstructured web content, a more challenging DR task should additionally utilize structured knowledge to provide a solid data foundation, facilitate quantitative computation, and lead to in-depth analyses. In this paper, we refer to this novel task as Knowledgeable Deep Research (KDR), which requires DR agents to generate reports with both structured and unstructured knowledge. Furthermore, we propose the Hybrid Knowledge Analysis framework (HKA), a multi-agent architecture that reasons over both kinds of knowledge and integrates the texts, figures, and tables into coherent multimodal reports. The key design is the Structured Knowledge Analyzer, which utilizes both coding and vision-language models to produce figures, tables, and corresponding insights. To support systematic evaluation, we construct KDR-Bench, which covers 9 domains, includes 41 expert-level questions, and incorporates a large number of structured knowledge resources (e.g., 1,252 tables). We further annotate the main conclusions and key points for each question and propose three categories of evaluation metrics including general-purpose, knowledge-centric, and vision-enhanced ones. Experimental results demonstrate that HKA consistently outperforms most existing DR agents on general-purpose and knowledge-centric metrics, and even surpasses the Gemini DR agent on vision-enhanced metrics, highlighting its effectiveness in deep, structure-aware knowledge analysis. Finally, we hope this work can serve as a new foundation for structured knowledge analysis in DR agents and facilitate future multimodal DR studies.

Yingjie Tang, Zixuan Li, Miloš Hašan et al.

Digitizing woven fabrics would be valuable for many applications, from digital humans to interior design. Previous work introduces a lightweight woven fabric acquisition approach by capturing a single reflection image and estimating the fabric parameters with a differentiable geometric and shading model. The renderings of the estimated fabric parameters can closely match the photo; however, the captured reflection image is insufficient to fully characterize the fabric sample reflectance. For instance, fabrics with different thicknesses might have similar reflection images but lead to significantly different transmission. We propose to recover the woven fabric parameters from two captured images: reflection and transmission. At the core of our method is a differentiable bidirectional scattering distribution function (BSDF) model, handling reflection and transmission, including single and multiple scattering. We propose a two-layer model, where the single scattering uses an SGGX phase function as in previous work, and multiple scattering uses a new azimuthally-invariant microflake definition, which we term ASGGX. This new fabric BSDF model closely matches real woven fabrics in both reflection and transmission. We use a simple setup for capturing reflection and transmission photos with a cell phone camera and two point lights, and estimate the fabric parameters via a lightweight network, together with a differentiable optimization. We also model the out-of-focus effects explicitly with a simple solution to match the thin-lens camera better. As a result, the renderings of the estimated parameters can agree with the input images on both reflection and transmission for the first time. The code for this paper is at https://github.com/lxtyin/FabricBTDF-Recovery.

Yuxin Zuo, Wenxuan Jiang, Wenxuan Liu et al. · bytedance

Empirical evidence indicates that LLMs exhibit spontaneous cross-lingual alignment. However, although LLMs show promising cross-lingual alignment in Information Extraction (IE), a significant imbalance across languages persists, highlighting an underlying deficiency. To address this, we propose KnowCoder-X, a powerful code LLM with advanced cross-lingual and multilingual capabilities for universal IE. Firstly, it standardizes the representation of multilingual schemas using Python classes, ensuring a consistent ontology across different languages. Then, IE across languages is formulated as a unified code generation task. Secondly, we conduct IE cross-lingual alignment instruction tuning on the translated instance prediction task to enhance the model's cross-lingual transferability. During this phase, we also construct a high-quality and diverse bilingual IE parallel dataset with 257k samples, called ParallelNER, synthesized by our proposed robust three-stage pipeline, with manual annotation to ensure quality. Although without training in 29 unseen languages, KnowCoder-X surpasses ChatGPT by 30.17\% and SoTA by 20.03\%, thereby demonstrating superior cross-lingual IE capabilities. Comprehensive evaluations on 64 IE benchmarks in Chinese and English under various settings demonstrate that KnowCoder-X significantly enhances cross-lingual IE transfer through boosting the IE alignment. Our code and dataset are available at: https://github.com/ICT-GoKnow/KnowCoder

Shujun Xia, Haokun Lin, Yichen Wu et al.

LLMs hold great promise for healthcare applications, but the rapid evolution of medical knowledge and errors in training data often cause them to generate outdated or inaccurate information, limiting their applicability in high-stakes clinical practice. Model editing has emerged as a potential remedy without full retraining. While parameter-based editing often compromises locality and is thus ill-suited for the medical domain, retrieval-based editing offers a more viable alternative. However, it still faces two critical challenges: (1) representation overlap within the medical knowledge space often causes inaccurate retrieval and reduces editing accuracy; (2) existing methods are restricted to single-sample edits, while batch-editing remains largely unexplored despite its importance for real-world medical applications. To address these challenges, we first construct MedVersa, an enhanced benchmark with broader coverage of medical subjects, designed to evaluate both single and batch edits under strict locality constraints. We then propose MedREK, a retrieval-based editing framework that integrates a shared query-key module for precise matching with an attention-based prompt encoder for informative guidance. Experimental results on various medical benchmarks demonstrate that our MedREK achieves superior performance across different core metrics and provides the first validated solution for batch-editing in medical LLMs. Our code and dataset are available at https://github.com/mylittleriver/MedREK.

Weichu Liu, Jing Xiong, Yuxuan Hu et al.

Large language models (LLMs) make significant progress in Emotional Intelligence (EI) and long-context understanding. However, existing benchmarks tend to overlook certain aspects of EI in long-context scenarios, especially under realistic, practical settings where interactions are lengthy, diverse, and often noisy. To move towards such realistic settings, we present LongEmotion, a benchmark specifically designed for long-context EI tasks. It covers a diverse set of tasks, including Emotion Classification, Emotion Detection, Emotion QA, Emotion Conversation, Emotion Summary, and Emotion Expression. On average, the input length for these tasks reaches 8,777 tokens, with long-form generation required for Emotion Expression. To enhance performance under realistic constraints, we incorporate Retrieval-Augmented Generation (RAG) and Collaborative Emotional Modeling (CoEM), and compare them with standard prompt-based methods. Unlike conventional approaches, our RAG method leverages both the conversation context and the large language model itself as retrieval sources, avoiding reliance on external knowledge bases. The CoEM method further improves performance by decomposing the task into five stages, integrating both retrieval augmentation and limited knowledge injection. Experimental results show that both RAG and CoEM consistently enhance EI-related performance across most long-context tasks, advancing LLMs toward more practical and real-world EI applications. Furthermore, we conducted a comparative case study experiment on the GPT series to demonstrate the differences among various models in terms of EI. Code is available on GitHub at https://github.com/LongEmotion/LongEmotion, and the project page can be found at https://longemotion.github.io/.

Zixuan Li, Binzong Geng, Jing Xiong et al.

Click-Through Rate (CTR) prediction, a core task in recommendation systems, estimates user click likelihood using historical behavioral data. Modeling user behavior sequences as text to leverage Language Models (LMs) for this task has gained traction, owing to LMs' strong semantic understanding and contextual modeling capabilities. However, a critical structural gap exists: user behavior sequences consist of discrete actions connected by semantically empty separators, differing fundamentally from the coherent natural language in LM pre-training. This mismatch causes semantic fragmentation, where LM attention scatters across irrelevant tokens instead of focusing on meaningful behavior boundaries and inter-behavior relationships, degrading prediction performance. To address this, we propose $\textit{CTR-Sink}$, a novel framework introducing behavior-level attention sinks tailored for recommendation scenarios. Inspired by attention sink theory, it constructs attention focus sinks and dynamically regulates attention aggregation via external information. Specifically, we insert sink tokens between consecutive behaviors, incorporating recommendation-specific signals such as temporal distance to serve as stable attention sinks. To enhance generality, we design a two-stage training strategy that explicitly guides LM attention toward sink tokens and a attention sink mechanism that amplifies inter-sink dependencies to better capture behavioral correlations. Experiments on one industrial dataset and two open-source datasets (MovieLens, Kuairec), alongside visualization results, validate the method's effectiveness across scenarios.

Hongyuan Lu, Zixuan Li, Wai Lam

As current training data for Large Language Models (LLMs) are dominated by English corpus, they are English-centric and they present impressive performance on English reasoning tasks.\footnote{This paper primarily studies English-centric models, but our method could be universal by using the centric language in the dictionary for non-English-centric LLMs.} Yet, they usually suffer from lower performance in other languages. There are about 7,000 languages over the world, and many are low-resourced on English-centric LLMs. For the sake of people who primarily speak these languages, it is especially urgent to enable our LLMs in those languages. Model training is usually effective, but computationally expensive and requires experienced NLP practitioners. This paper presents a novel and simple yet effective method called \textbf{D}ictionary \textbf{I}nsertion \textbf{P}rompting (\textbf{DIP}). When providing a non-English prompt, DIP looks up a word dictionary and inserts words' English counterparts into the prompt for LLMs. It then enables better translation into English and better English model thinking steps which leads to obviously better results. We experiment with about 200 languages from FLORES-200. Since there are no adequate datasets, we use the NLLB translator to create synthetic multilingual benchmarks from the existing 4 English reasoning benchmarks such as GSM8K and AQuA. Despite the simplicity and computationally lightweight, we surprisingly found the effectiveness of DIP on math and commonsense reasoning tasks on multiple open-source and close-source LLMs.\footnote{Our dictionaries, code, and synthetic benchmarks will be open-sourced to facilitate future research.}

Jing Xiong, Zixuan Li, Chuanyang Zheng et al.

Recent advances in natural language processing, primarily propelled by Large Language Models (LLMs), have showcased their remarkable capabilities grounded in in-context learning. A promising avenue for guiding LLMs in intricate reasoning tasks involves the utilization of intermediate reasoning steps within the Chain-of-Thought (CoT) paradigm. Nevertheless, the central challenge lies in the effective selection of exemplars for facilitating in-context learning. In this study, we introduce a framework that leverages Dual Queries and Low-rank approximation Re-ranking (DQ-LoRe) to automatically select exemplars for in-context learning. Dual Queries first query LLM to obtain LLM-generated knowledge such as CoT, then query the retriever to obtain the final exemplars via both question and the knowledge. Moreover, for the second query, LoRe employs dimensionality reduction techniques to refine exemplar selection, ensuring close alignment with the input question's knowledge. Through extensive experiments, we demonstrate that DQ-LoRe significantly outperforms prior state-of-the-art methods in the automatic selection of exemplars for GPT-4, enhancing performance from 92.5% to 94.2%. Our comprehensive analysis further reveals that DQ-LoRe consistently outperforms retrieval-based approaches in terms of both performance and adaptability, especially in scenarios characterized by distribution shifts. DQ-LoRe pushes the boundary of in-context learning and opens up new avenues for addressing complex reasoning challenges. Our code is released at https://github.com/menik1126/DQ-LoRe

Zixuan Li, Yutao Zeng, Yuxin Zuo et al. · bytedance

In this paper, we propose KnowCoder, a Large Language Model (LLM) to conduct Universal Information Extraction (UIE) via code generation. KnowCoder aims to develop a kind of unified schema representation that LLMs can easily understand and an effective learning framework that encourages LLMs to follow schemas and extract structured knowledge accurately. To achieve these, KnowCoder introduces a code-style schema representation method to uniformly transform different schemas into Python classes, with which complex schema information, such as constraints among tasks in UIE, can be captured in an LLM-friendly manner. We further construct a code-style schema library covering over $\textbf{30,000}$ types of knowledge, which is the largest one for UIE, to the best of our knowledge. To ease the learning process of LLMs, KnowCoder contains a two-phase learning framework that enhances its schema understanding ability via code pretraining and its schema following ability via instruction tuning. After code pretraining on around $1.5$B automatically constructed data, KnowCoder already attains remarkable generalization ability and achieves relative improvements by $\textbf{49.8%}$ F1, compared to LLaMA2, under the few-shot setting. After instruction tuning, KnowCoder further exhibits strong generalization ability on unseen schemas and achieves up to $\textbf{12.5%}$ and $\textbf{21.9%}$, compared to sota baselines, under the zero-shot setting and the low resource setting, respectively. Additionally, based on our unified schema representations, various human-annotated datasets can simultaneously be utilized to refine KnowCoder, which achieves significant improvements up to $\textbf{7.5%}$ under the supervised setting.

Fang Long, Wenkang Su, Zixuan Li et al.

Adverse weather removal aims to restore clear vision under adverse weather conditions. Existing methods are mostly tailored for specific weather types and rely heavily on extensive labeled data. In dealing with these two limitations, this paper presents a pioneering semi-supervised all-in-one adverse weather removal framework built on the teacher-student network with a Denoising Diffusion Model (DDM) as the backbone, termed SemiDDM-Weather. As for the design of DDM backbone in our SemiDDM-Weather, we adopt the SOTA Wavelet Diffusion Model-Wavediff with customized inputs and loss functions, devoted to facilitating the learning of many-to-one mapping distributions for efficient all-in-one adverse weather removal with limited label data. To mitigate the risk of misleading model training due to potentially inaccurate pseudo-labels generated by the teacher network in semi-supervised learning, we introduce quality assessment and content consistency constraints to screen the "optimal" outputs from the teacher network as the pseudo-labels, thus more effectively guiding the student network training with unlabeled data. Experimental results show that on both synthetic and real-world datasets, our SemiDDM-Weather consistently delivers high visual quality and superior adverse weather removal, even when compared to fully supervised competitors. Our code and pre-trained model are available at this repository.

Yuru Jing, Zixuan Li

A chest X-ray is one of the most widely available radiological examinations for diagnosing and detecting various lung illnesses. The National Institutes of Health (NIH) provides an extensive database, ChestX-ray8 and ChestXray14, to help establish a deep learning community for analysing and predicting lung diseases. ChestX-ray14 consists of 112,120 frontal-view X-ray images of 30,805 distinct patients with text-mined fourteen disease image labels, where each image has multiple labels and has been utilised in numerous research in the past. To our current knowledge, no previous study has investigated outliers and multi-label impact for a single X-ray image during the preprocessing stage. The effect of outliers is mitigated in this paper by our proposed auto-outlier fusion technique. The image label is regenerated by concentrating on a particular factor in one image. The final cleaned dataset will be used to compare the mechanisms of multi-head self-attention and multi-head attention with generalised max-pooling.

Xuhui Jiang, Yinghan Shen, Zhichao Shi et al.

Entity Alignment (EA) is vital for integrating diverse knowledge graph (KG) data, playing a crucial role in data-driven AI applications. Traditional EA methods primarily rely on comparing entity embeddings, but their effectiveness is constrained by the limited input KG data and the capabilities of the representation learning techniques. Against this backdrop, we introduce ChatEA, an innovative framework that incorporates large language models (LLMs) to improve EA. To address the constraints of limited input KG data, ChatEA introduces a KG-code translation module that translates KG structures into a format understandable by LLMs, thereby allowing LLMs to utilize their extensive background knowledge to improve EA accuracy. To overcome the over-reliance on entity embedding comparisons, ChatEA implements a two-stage EA strategy that capitalizes on LLMs' capability for multi-step reasoning in a dialogue format, thereby enhancing accuracy while preserving efficiency. Our experimental results verify ChatEA's superior performance, highlighting LLMs' potential in facilitating EA tasks.

Zhuo Chen, Zhao Zhang, Zixuan Li et al. · bytedance

Temporal Knowledge Graph Question Answering (TKGQA) aims to answer questions with temporal intent over Temporal Knowledge Graphs (TKGs). The core challenge of this task lies in understanding the complex semantic information regarding multiple types of time constraints (e.g., before, first) in questions. Existing end-to-end methods implicitly model the time constraints by learning time-aware embeddings of questions and candidate answers, which is far from understanding the question comprehensively. Motivated by semantic-parsing-based approaches that explicitly model constraints in questions by generating logical forms with symbolic operators, we design fundamental temporal operators for time constraints and introduce a novel self-improvement Programming method for TKGQA (Prog-TQA). Specifically, Prog-TQA leverages the in-context learning ability of Large Language Models (LLMs) to understand the combinatory time constraints in the questions and generate corresponding program drafts with a few examples given. Then, it aligns these drafts to TKGs with the linking module and subsequently executes them to generate the answers. To enhance the ability to understand questions, Prog-TQA is further equipped with a self-improvement strategy to effectively bootstrap LLMs using high-quality self-generated drafts. Extensive experiments demonstrate the superiority of the proposed Prog-TQA on MultiTQ and CronQuestions datasets, especially in the Hits@1 metric.

Ye Yuan, Youyuan Zhang, Can Chen et al.

Offline model-based optimization (MBO) aims to maximize a black-box objective function using only an offline dataset of designs and scores. These tasks span various domains, such as robotics, material design, and protein and molecular engineering. A common approach involves training a surrogate model using existing designs and their corresponding scores, and then generating new designs through gradient-based updates with respect to the surrogate model. This method suffers from the out-of-distribution issue, where the surrogate model may erroneously predict high scores for unseen designs. To address this challenge, we introduce a novel method, Design Editing for Offline Model-based Optimization (DEMO), which leverages a diffusion prior to calibrate overly optimized designs. DEMO first generates pseudo design candidates by performing gradient ascent with respect to a surrogate model. While these pseudo design candidates contain information beyond the offline dataset, they might be invalid or have erroneously high predicted scores. Therefore, to address this challenge while utilizing the information provided by pseudo design candidates, we propose an editing process to refine these pseudo design candidates. We introduce noise to the pseudo design candidates and subsequently denoise them with a diffusion prior trained on the offline dataset, ensuring they align with the distribution of valid designs. Empirical evaluations on seven offline MBO tasks show that, with properly tuned hyperparameters, DEMOs score is competitive with the best previously reported scores in the literature.

Tianxing Chen, Kaixuan Wang, Zhaohui Yang et al.

Embodied Artificial Intelligence (Embodied AI) is an emerging frontier in robotics, driven by the need for autonomous systems that can perceive, reason, and act in complex physical environments. While single-arm systems have shown strong task performance, collaborative dual-arm systems are essential for handling more intricate tasks involving rigid, deformable, and tactile-sensitive objects. To advance this goal, we launched the RoboTwin Dual-Arm Collaboration Challenge at the 2nd MEIS Workshop, CVPR 2025. Built on the RoboTwin Simulation platform (1.0 and 2.0) and the AgileX COBOT-Magic Robot platform, the competition consisted of three stages: Simulation Round 1, Simulation Round 2, and a final Real-World Round. Participants totally tackled 17 dual-arm manipulation tasks, covering rigid, deformable, and tactile-based scenarios. The challenge attracted 64 global teams and over 400 participants, producing top-performing solutions like SEM and AnchorDP3 and generating valuable insights into generalizable bimanual policy learning. This report outlines the competition setup, task design, evaluation methodology, key findings and future direction, aiming to support future research on robust and generalizable bimanual manipulation policies. The Challenge Webpage is available at https://robotwin-benchmark.github.io/cvpr-2025-challenge/.

Wenxuan Liu, Zixuan Li, Long Bai et al.

Developing a general-purpose extraction system that can extract events with massive types is a long-standing target in Event Extraction (EE). In doing so, the challenge comes from two aspects: 1) The absence of an efficient and effective annotation method. 2) The absence of a powerful extraction method can handle massive types. For the first challenge, we propose a collaborative annotation method based on Large Language Models (LLMs). Through collaboration among multiple LLMs, it first refines annotations of trigger words from distant supervision and then carries out argument annotation. Next, a voting phase consolidates the annotation preferences across different LLMs. Finally, we create the EEMT dataset, the largest EE dataset to date, featuring over 200,000 samples, 3,465 event types, and 6,297 role types. For the second challenge, we propose an LLM-based Partitioning EE method called LLM-PEE. To overcome the limited context length of LLMs, LLM-PEE first recalls candidate event types and then splits them into multiple partitions for LLMs to extract events. The results in the supervised setting show that LLM-PEE outperforms the state-of-the-art methods by 5.4 in event detection and 6.1 in argument extraction. In the zero-shot setting, LLM-PEE achieves up to 12.9 improvement compared to mainstream LLMs, demonstrating its strong generalization capabilities.

Liumeng Xue, Ziya Zhou, Jiahao Pan et al.

Recent advancements in audio tokenization have significantly enhanced the integration of audio capabilities into large language models (LLMs). However, audio understanding and generation are often treated as distinct tasks, hindering the development of truly unified audio-language models. While instruction tuning has demonstrated remarkable success in improving generalization and zero-shot learning across text and vision, its application to audio remains largely unexplored. A major obstacle is the lack of comprehensive datasets that unify audio understanding and generation. To address this, we introduce Audio-FLAN, a large-scale instruction-tuning dataset covering 80 diverse tasks across speech, music, and sound domains, with over 100 million instances. Audio-FLAN lays the foundation for unified audio-language models that can seamlessly handle both understanding (e.g., transcription, comprehension) and generation (e.g., speech, music, sound) tasks across a wide range of audio domains in a zero-shot manner. The Audio-FLAN dataset is available on HuggingFace and GitHub and will be continuously updated.

Xiaoshuo Yan, Zhaochuan Li, Lei Meng et al.

Causal inference has emerged as a promising approach to mitigate long-tail classification by handling the biases introduced by class imbalance. However, along with the change of advanced backbone models from Convolutional Neural Networks (CNNs) to Visual Transformers (ViT), existing causal models may not achieve an expected performance gain. This paper investigates the influence of existing causal models on CNNs and ViT variants, highlighting that ViT's global feature representation makes it hard for causal methods to model associations between fine-grained features and predictions, which leads to difficulties in classifying tail classes with similar visual appearance. To address these issues, this paper proposes TSCNet, a two-stage causal modeling method to discover fine-grained causal associations through multi-scale causal interventions. Specifically, in the hierarchical causal representation learning stage (HCRL), it decouples the background and objects, applying backdoor interventions at both the patch and feature level to prevent model from using class-irrelevant areas to infer labels which enhances fine-grained causal representation. In the counterfactual logits bias calibration stage (CLBC), it refines the optimization of model's decision boundary by adaptive constructing counterfactual balanced data distribution to remove the spurious associations in the logits caused by data distribution. Extensive experiments conducted on various long-tail benchmarks demonstrate that the proposed TSCNet can eliminate multiple biases introduced by data imbalance, which outperforms existing methods.

Yude Zou, Junji Gong, Xing Gao et al.

Human-object-scene interactions (HOSI) generation has broad applications in embodied AI, simulation, and animation. Unlike human-object interaction (HOI) and human-scene interaction (HSI), HOSI generation requires reasoning over dynamic object-scene changes, yet suffers from limited annotated data. To address these issues, we propose a coarse-to-fine instruction-conditioned interaction generation framework that is explicitly aligned with the iterative denoising process of a consistency model. In particular, we adopt a dynamic perception strategy that leverages trajectories from the preceding refinement to update scene context and condition subsequent refinement at each denoising step of consistency model, yielding consistent interactions. To further reduce physical artifacts, we introduce a bump-aware guidance that mitigates collisions and penetrations during sampling without requiring fine-grained scene geometry, enabling real-time generation. To overcome data scarcity, we design a hybrid training startegy that synthesizes pseudo-HOSI samples by injecting voxelized scene occupancy into HOI datasets and jointly trains with high-fidelity HSI data, allowing interaction learning while preserving realistic scene awareness. Extensive experiments demonstrate that our method achieves state-of-the-art performance in both HOSI and HOI generation, and strong generalization to unseen scenes. Project page: https://yudezou.github.io/InfBaGel-page/

Thierry Blankenstein, Jialin Yu, Zixuan Li et al.

Agents backed by large language models (LLMs) often rely on external tools drawn from marketplaces where multiple providers offer functionally equivalent options. This raises a critical point concerning fairness: if selection is systematically biased, it can degrade user experience and distort competition by privileging some providers over others. We introduce a benchmark of diverse tool categories, each containing multiple functionally equivalent tools, to evaluate tool-selection bias. Using this benchmark, we test seven models and show that unfairness exists with models either fixating on a single provider or disproportionately preferring earlier-listed tools in context. To investigate the origins of this bias, we conduct controlled experiments examining tool features, metadata (name, description, parameters), and pre-training exposure. We find that: (1) semantic alignment between queries and metadata is the strongest predictor of choice; (2) perturbing descriptions significantly shifts selections; and (3) repeated pre-training exposure to a single endpoint amplifies bias. Finally, we propose a lightweight mitigation that first filters the candidate tools to a relevant subset and then samples uniformly, reducing bias while preserving good task coverage. Our findings highlight tool-selection bias as a key obstacle for the fair deployment of tool-augmented LLMs.

Long Bai, Zixuan Li, Xiaolong Jin et al.

Forecasting over Temporal Knowledge Graphs (TKGs) which predicts future facts based on historical ones has received much attention. Recent studies have introduced Large Language Models (LLMs) for this task to enhance the models' generalization abilities. However, these models perform forecasting via simultaneously learning two kinds of entangled knowledge in the TKG: (1) general patterns, i.e., invariant temporal structures shared across different scenarios; and (2) scenario information, i.e., factual knowledge engaged in specific scenario, such as entities and relations. As a result, the learning processes of these two kinds of knowledge may interfere with each other, which potentially impact the generalization abilities of the models. To enhance the generalization ability of LLMs on this task, in this paper, we propose a General-to-Specific learning framework (G2S) that disentangles the learning processes of the above two kinds of knowledge. In the general learning stage, we mask the scenario information in different TKGs and convert it into anonymous temporal structures. After training on these structures, the model is able to capture the general patterns across different TKGs. In the specific learning stage, we inject the scenario information into the structures via either in-context learning or fine-tuning modes. Experimental results show that G2S effectively improves the generalization abilities of LLMs.

Jizhao Zhu, Akang Shi, Zixuan Li et al.

In this paper, we aim to enhance the robustness of Universal Information Extraction (UIE) by introducing a new benchmark dataset, a comprehensive evaluation, and a feasible solution. Existing robust benchmark datasets have two key limitations: 1) They generate only a limited range of perturbations for a single Information Extraction (IE) task, which fails to evaluate the robustness of UIE models effectively; 2) They rely on small models or handcrafted rules to generate perturbations, often resulting in unnatural adversarial examples. Considering the powerful generation capabilities of Large Language Models (LLMs), we introduce a new benchmark dataset for Robust UIE, called RUIE-Bench, which utilizes LLMs to generate more diverse and realistic perturbations across different IE tasks. Based on this dataset, we comprehensively evaluate existing UIE models and reveal that both LLM-based models and other models suffer from significant performance drops. To improve robustness and reduce training costs, we propose a data-augmentation solution that dynamically selects hard samples for iterative training based on the model's inference loss. Experimental results show that training with only \textbf{15\%} of the data leads to an average \textbf{7.5\%} relative performance improvement across three IE tasks.

Zhuo Chen, Fei Wang, Zixuan Li et al.

Knowledge Base Question Answering (KBQA) aims to answer natural-language questions over a structured Knowledge Base (KB). Recent work improves KBQA by adopting an agentic reasoning paradigm, in which Large Language Models (LLMs) iteratively decompose a question, generate its corresponding logical queries, and interact with the KB to derive the answer. However, these methods typically fine-tune LLMs on reasoning trajectories synthesized via process supervision, which offers weak incentives for exploration and thus fails to strengthen the agentic reasoning ability. In this paper, we propose KnowCoder-A1, an LLM that can autonomously perform agentic reasoning on KBs to obtain answers. To incentivize autonomous exploration, KnowCoder-A1 trains the LLM under outcome-only supervision via a multi-stage curriculum reinforcement learning with an easy-to-hard curriculum. To establish foundational agentic capabilities, KnowCoder-A1 first fine-tunes the LLM on a small set of high-quality trajectories obtained through outcome-based rejection sampling. Then, to alleviate the reward sparsity inherent in outcome-only supervision, it applies multi-stage curriculum RL with reward schedules that progress from easy to hard. Trained with outcome-only supervision, KnowCoder-A1 exhibits powerful reasoning behaviors and consistently outperforms prior approaches across three mainstream datasets. Notably, on the zero-shot subset of GrailQA, KnowCoder-A1 achieves up to an 11.1% relative improvement while using only one-twelfth of the training data, demonstrating strong agentic reasoning capabilities.