Haotian Xu, Yue Hu, Zhengqiu Zhu, Chen Gao, Ziyou Wang, Junreng Rao, Wenhao Lu, Weishi Li, Quanjun Yin, Yong Li. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Baining Zhao, Jianjie Fang, Zichao Dai, Ziyou Wang, Jirong Zha, Weichen Zhang, Chen Gao, Yue Wang, Jinqiang Cui, Xinlei Chen, Yong Li. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

How to enable agents to predict the outcomes of their own motion intentions in three-dimensional space has been a fundamental problem in embodied intelligence. To explore general spatial imagination capability, we present AirScape, the first world model designed for six-degree-of-freedom aerial agents. AirScape predicts future observation sequences based on current visual inputs and motion intentions. Specifically, we construct a dataset for aerial world model training and testing, which consists of 11k video-intention pairs. This dataset includes first-person-view videos capturing diverse drone actions across a wide range of scenarios, with over 1,000 hours spent annotating the corresponding motion intentions. Then we develop a two-phase schedule to train a foundation model-initially devoid of embodied spatial knowledge-into a world model that is controllable by motion intentions and adheres to physical spatio-temporal constraints. Experimental results demonstrate that AirScape significantly outperforms existing foundation models in 3D spatial imagination capabilities, especially with over a 50% improvement in metrics reflecting motion alignment. The project is available at: https://embodiedcity.github.io/AirScape/.

Humans can perceive and reason about spatial relationships from sequential visual observations, such as egocentric video streams. However, how pretrained models acquire such abilities, especially high-level reasoning, remains unclear. This paper introduces Embodied-R, a collaborative framework combining large-scale Vision-Language Models (VLMs) for perception and small-scale Language Models (LMs) for reasoning. Using Reinforcement Learning (RL) with a novel reward system considering think-answer logical consistency, the model achieves slow-thinking capabilities with limited computational resources. After training on only 5k embodied video samples, Embodied-R with a 3B LM matches state-of-the-art multimodal reasoning models (OpenAI-o1, Gemini-2.5-pro) on both in-distribution and out-of-distribution embodied spatial reasoning tasks. Embodied-R also exhibits emergent thinking patterns such as systematic analysis and contextual integration. We further explore research questions including response length, training on VLM, strategies for reward design, and differences in model generalization after SFT (Supervised Fine-Tuning) and RL training. The project page is available at: https://embodiedcity.github.io/Embodied-R/.

While people get used to surfing web, managing the overload of web applications has become a critical problem for application providers. Targeting overload issue of complex, dynamic web applications, this paper presents an adaptive service degradation approach. Our approach attempts to automatically locate the bottleneck inside the applications and generate proper degradation plans in real time. This is accomplished through internally monitoring the performance and resource utilization state of the application, which is decomposed into a set of services. By dynamic controlling the bottleneck of resource utilization, the application can keep providing key services even when overload occurs, for example, by degrading only the service which consumes critical resources and has a low priority from the perspective of business logic. We implement a prototype and conduct a case study with a business application. The case study demonstrates the approach is effective when overload occurs.

In order to support EJB3.0, many open-source application server vendors have implemented a new container which is independent with the existing EJB2.x one. However, the functional redundancy between both of the containers increases the maintenance cost when the software are modified because of problems or new requirements. This paper proposes a Dual-container to support EJB2.x and EJB3.0 in a unified architecture. Software engineering perspective is applied, three principles including greatest-common-divisor, lease-common-multiple, and transparent intrusion, are presented to guide the construction of the architecture. We implemented the Dual-container in an open-source J2EE application server. The evaluation results show that the Dual-container is effective in reducing the functional redundancy and increasing the maintainability. We demonstrate that in a comparative analysis.