Legal reasoning is a fundamental component of legal analysis and decision-making. Existing computational approaches to legal reasoning predominantly rely on generic reasoning frameworks such as syllogism, which do not comprehensively examine the nuanced process of legal reasoning. Moreover, current research has largely focused on criminal cases, with insufficient modeling for civil cases. In this work, we present a novel framework to explicitly model legal reasoning in the analysis of Chinese tort-related civil cases. We first operationalize the legal reasoning process in tort analysis into the three-module LexChain framework, with each module consisting of multiple finer-grained sub-steps. Informed by the LexChain framework, we introduce the task of tort legal reasoning and construct an evaluation benchmark to systematically assess the critical steps within analytical reasoning chains for tort analysis. Leveraging this benchmark, we evaluate existing large language models for their legal reasoning ability in civil tort contexts. Our results indicate that current models still fall short in accurately handling crucial elements of tort legal reasoning. Furthermore, we introduce several baseline approaches that explicitly incorporate LexChain-style reasoning through prompting or post-training. The proposed baselines achieve significant improvements in tort-related legal reasoning and generalize well to related legal analysis tasks, demonstrating the value of explicitly modeling legal reasoning chains to enhance the reasoning capabilities of language models.

Yida Cai, Ranjuexiao Hu, Huiyuan Xie, Chenyang Li, Yun Liu, Yuxiao Ye, Zhenghao Liu, Weixing Shen, Zhiyuan Liu. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Hao Wang, Hao Gu, Hongming Piao, Kaixiong Gong, Yuxiao Ye, Xiangyu Yue, Sirui Han, Yike Guo, Dapeng Wu. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Legal text generated by large language models (LLMs) can usually achieve reasonable factual accuracy, but it frequently fails to adhere to the specialised stylistic norms and linguistic conventions of legal writing. In order to improve stylistic quality, a crucial first step is to establish a reliable evaluation method. However, having legal experts manually develop such a metric is impractical, as the implicit stylistic requirements in legal writing practice are difficult to formalise into explicit rubrics. Meanwhile, existing automatic evaluation methods also fall short: reference-based metrics conflate semantic accuracy with stylistic fidelity, and LLM-as-a-judge evaluations suffer from opacity and inconsistency. To address these challenges, we introduce CLASE (Chinese LegAlese Stylistic Evaluation), a hybrid evaluation method that focuses on the stylistic performance of legal text. The method incorporates a hybrid scoring mechanism that combines 1) linguistic feature-based scores and 2) experience-guided LLM-as-a-judge scores. Both the feature coefficients and the LLM scoring experiences are learned from contrastive pairs of authentic legal documents and their LLM-restored counterparts. This hybrid design captures both surface-level features and implicit stylistic norms in a transparent, reference-free manner. Experiments on 200 Chinese legal documents show that CLASE achieves substantially higher alignment with human judgments than traditional metrics and pure LLM-as-a-judge methods. Beyond improved alignment, CLASE provides interpretable score breakdowns and suggestions for improvements, offering a scalable and practical solution for professional stylistic evaluation in legal text generation (Code and data for CLASE is available at: https://github.com/rexera/CLASE).

Real-world enterprise text-to-SQL workflows often involve complex cloud or local data across various database systems, multiple SQL queries in various dialects, and diverse operations from data transformation to analytics. We introduce Spider 2.0, an evaluation framework comprising $632$ real-world text-to-SQL workflow problems derived from enterprise-level database use cases. The databases in Spider 2.0 are sourced from real data applications, often containing over 1,000 columns and stored in local or cloud database systems such as BigQuery and Snowflake. We show that solving problems in Spider 2.0 frequently requires understanding and searching through database metadata, dialect documentation, and even project-level codebases. This challenge calls for models to interact with complex SQL workflow environments, process extremely long contexts, perform intricate reasoning, and generate multiple SQL queries with diverse operations, often exceeding $100$ lines, which goes far beyond traditional text-to-SQL challenges. Our evaluations indicate that based on o1-preview, our code agent framework successfully solves only 21.3\% of the tasks, compared with 91.2\% on Spider 1.0 and 73.0\% on BIRD. Our results on Spider 2.0 show that while language models have demonstrated remarkable performance in code generation --- especially in prior text-to-SQL benchmarks --- they require significant improvement in order to achieve adequate performance for real-world enterprise usage. Progress on Spider 2.0 represents crucial steps towards developing intelligent, autonomous, code agents for real-world enterprise settings. Our code, baseline models, and data are available at [spider2-sql.github.io](spider2-sql.github.io) .

Vision-language models have demonstrated impressive capabilities as computer-use agents (CUAs) capable of automating diverse computer tasks. As their commercial potential grows, critical details of the most capable CUA systems remain closed. As these agents will increasingly mediate digital interactions and execute consequential decisions on our behalf, the research community needs access to open CUA frameworks to study their capabilities, limitations, and risks. To bridge this gap, we propose OpenCUA, a comprehensive open-source framework for scaling CUA data and foundation models. Our framework consists of: (1) an annotation infrastructure that seamlessly captures human computer-use demonstrations; (2) AgentNet, the first large-scale computer-use task dataset spanning 3 operating systems and 200+ applications and websites; (3) a scalable pipeline that transforms demonstrations into state–action pairs with reflective long Chain-of-Thought reasoning that sustain robust performance gains as data scales. Our end-to-end agent models demonstrate strong performance across CUA benchmarks. In particular, OpenCUA-72B achieves an average success rate of 45.0% on OSWorld‑Verified, establishing a new state-of-the-art (SOTA) among open-source models. Further analysis confirms that our approach generalizes well across domains and benefits significantly from increased test-time computation. We release our annotation tool, datasets, code, and models to build open foundations for further CUA research.

By harnessing the capabilities of unmanned aerial and ground vehicles (UAVs and UGVs), equipped with high-precision sensors, air-ground mobile crowdsensing (AG-MCS) has proven to be effective for data collection in urban environments. In this paper, by optimizing the metric of age-of-information (AoI) that measures the freshness of collected data, we consider the problem of AoI-Aware AG-MCS (A3G-MCS), where UGVs dispatch UAVs from multiple UGV stops to collect data from point-of-interests (PoIs). We propose a novel multi-agent curriculum learning framework called “MACL(MCS)”, that explicitly balances the individual and team goals of both UAV/UGV controllers to facilitate the exploration of policy towards globally-optimal performance. It is further enhanced by a UAV/UGV collaborative observation augmentation (COA) module for improved inter-controller communication. Extensive results reveal that MACL(MCS) consistently outperforms five baselines, and achieves comparable performance to exact method with better scalability and efficiency. It also showcases strong generalization capability towards real-world scenarios on both TSPLIB and Purdue, KAIST and NCSU datasets.

Metaverse is expected to provide mobile users with emerging applications both in regular situation like intelligent transportation services and in emergencies like wireless search and disaster response. These applications are usually associated with stringent quality-of-information (QoI) requirements like throughput and age-of-information (AoI), which can be further guaranteed by using unmanned aerial vehicles (UAVs) as aerial base stations (BSs) to compensate the existing 5G infrastructures. In this paper, we consider a new QoI-aware mobile crowdsensing (MCS) campaign by UAVs which move around and collect data from mobile users wearing metaverse devices. Specifically, we propose “MetaCS”, a multi-agent deep reinforcement learning (MADRL) framework with improvements on a Transformer-based user mobility prediction module between regions and a relational graph learning mechanism to enable the selection of most informative partners to communicate for each UAV. Extensive results and trajectory visualizations on three real mobility datasets in NCSU, KAIST and Beijing show that MetaCS consistently outperforms six baselines in terms of overall QoI index, when varying different numbers of UAVs, throughput requirement, and AoI threshold.

Spatial crowdsourcing (SC) has proven as a promising paradigm to employ human workers to collect data from diverse Point-of-Interests (PoIs) in a given area. Different from using human participants, we propose a novel air-ground SC scenario to fully take advantage of benefits brought by unmanned vehicles (UVs), including unmanned aerial vehicles (UAVs) with controllable high mobility and unmanned ground vehicles (UGVs) with abundant sensing resources. The objective is to maximize the amount of collected data, geographical fairness among all PoIs, and minimize the data loss and energy consumption, integrated as one single metric called "efficiency". We explicitly explore both individuality and cooperation natures of UAVs and UGVs by proposing a multi-agent deep reinforcement learning (MADRL) framework called "h/i-MADRL". Compatible with all multi-agent actor-critic methods, h/i-MADRL adds two novel plug-in modules: (a) h-CoPO, which models the cooperation preference among heterogeneous UAVs and UGVs; and (b) i-EOI, which extracts the UV’s individuality and encourages a better spatial division of work by adding intrinsic reward. Extensive experimental results on two real-world datasets on Purdue and NCSU campuses confirm that h/i-MADRL achieves a better exploration of both individuality and cooperation simultaneously, resulting in a better performance in terms of efficiency compared with five baselines.

Mobile Crowdsensing (MCS) with smart devices has become an appealing paradigm for urban sensing. With the development of 5G-and-beyond technologies, unmanned aerial vehicles (UAVs) become possible for real-time applications, including wireless coverage, search and even disaster response. In this paper, we consider to use a group of UAVs as aerial base stations (BSs) to move around and collect data from multiple MCS users, forming a UAV crowdsensing campaign (UCS). Our goal is to maximize the collected data, geographical coverage whiling minimizing the age-of-information (AoI) of all mobile users simultaneously, with efficient use of constrained energy reserve. We propose a model-based deep reinforcement learning (DRL) framework called "GCRL-min(AoI)", which mainly consists of a novel model-based Monte Carlo tree search (MCTS) structure based on state-of-the-art approach MCTS (AlphaZero). We further improve it by adding a spatial UAV-user correlation extraction mechanism by a relational graph convolutional network (RGCN), and a next state prediction module to reduce the dependance of experience data. Extensive results and trajectory visualization on three real human mobility datasets in Purdue University, KAIST and NCSU show that GCRL-min(AoI) consistently outperforms five baselines, when varying different number of UAVs and maximum coupling loss in terms of four metrics.

Non-neural approaches to argument mining (AM) are often pipelined and require heavy feature-engineering.In this paper, we propose a neural end-to-end approach to AM which is based on dependency parsing, in contrast to the current state-of-the-art which relies on relation extraction.Our biaffine AM dependency parser significantly outperforms the state-ofthe-art, performing at F 1 = 73.5% for component identification and F 1 = 46.4% for relation identification.One of the advantages of treating AM as biaffine dependency parsing is the simple neural architecture that results.The idea of treating AM as dependency parsing is not new, but has previously been abandoned as it was lagging far behind the state-of-the-art.In a thorough analysis, we investigate the factors that contribute to the success of our model: the biaffine model itself, our representation for the dependency structure of arguments, different encoders in the biaffine model, and syntactic information additionally fed to the model.Our work demonstrates that dependency parsing for AM, an overlooked idea from the past, deserves more attention in the future.

Detecting user intents from utterances is the basis of natural language understanding (NLU) task. To understand the meaning of utterances, some work focuses on fully representing utterances via semantic parsing in which annotation cost is labor-intentsive. While some researchers simply view this as intent classification or frequently asked questions (FAQs) retrieval, they do not leverage the shared utterances among different intents. We propose a simple and novel multi-point semantic representation framework with relatively low annotation cost to leverage the fine-grained factor information, decomposing queries into four factors, i.e., topic, predicate, object/condition, query type. Besides, we propose a compositional intent bi-attention model under multi-task learning with three kinds of attention mechanisms among queries, labels and factors, which jointly combines coarse-grained intent and fine-grained factor information. Extensive experiments show that our framework and model significantly outperform several state-of-the-art approaches with an improvement of 1.35%-2.47% in terms of accuracy.