3D Semantic Occupancy Prediction is fundamental for spatial understanding, yet existing approaches face challenges in scalability and generalization due to their reliance on extensive labeled data and computationally intensive voxel-wise representations. In this paper, we introduce GaussTR, a novel Gaussian-based TRansformer framework that unifies sparse 3D modeling with foundation model alignment through Gaussian representations to advance 3D spatial understanding. GaussTR predicts sparse sets of Gaussians in a feed-forward manner to represent 3D scenes. By splatting the Gaussians into 2D views and aligning the rendered features with foundation models, GaussTR facilitates self-supervised 3D representation learning and enables open-vocabulary semantic occupancy prediction without requiring explicit annotations. Empirical experiments on the Occ3D-nuScenes dataset demonstrate GaussTR’s state-of-the-art zero-shot performance of 12.27 mIoU, along with a 40% reduction in training time. These results highlight the efficacy of GaussTR for scalable and holistic 3D spatial understanding, with promising implications in autonomous driving and embodied agents. The code is available at https://github.com/hustvl/GaussTR.

Large Vision-Language Models (LVLMs) have exhibited remarkable progress. However, deficiencies remain compared to human intelligence, such as hallucination and shallow pattern matching. In this work, we aim to evaluate a fundamental yet underexplored intelligence: association, a cornerstone of human cognition for creative thinking and knowledge integration. Current benchmarks, often limited to closed-ended tasks, fail to capture the complexity of open-ended association reasoning vital for real-world applications. To address this, we present MM-OPERA, a systematic benchmark with 11,497 instances across two open-ended tasks: Remote-Item Association (RIA) and In-Context Association (ICA), aligning association intelligence evaluation with human psychometric principles. It challenges LVLMs to resemble the spirit of divergent thinking and convergent associative reasoning through free-form responses and explicit reasoning paths. We deploy tailored LLM-as-a-Judge strategies to evaluate open-ended outputs, applying process-reward-informed judgment to dissect reasoning with precision. Extensive empirical studies on state-of-the-art LVLMs, including sensitivity analysis of task instances, validity analysis of LLM-as-a-Judge strategies, and diversity analysis across abilities, domains, languages, cultures, etc., provide a comprehensive and nuanced understanding of the limitations of current LVLMs in associative reasoning, paving the way for more human-like and general-purpose AI. The dataset and code are available at https://github.com/MM-OPERA-Bench/MM-OPERA.

We propose **GAM-Agent**, a game-theoretic multi-agent framework for enhancing vision-language reasoning. Unlike prior single-agent or monolithic models, GAM-Agent formulates the reasoning process as a non-zero-sum game between base agents—each specializing in visual perception subtasks—and a critical agent that verifies logic consistency and factual correctness. Agents communicate via structured claims, evidence, and uncertainty estimates. The framework introduces an uncertainty-aware controller to dynamically adjust agent collaboration, triggering multi-round debates when disagreement or ambiguity is detected. This process yields more robust and interpretable predictions. Experiments on four challenging benchmarks—MMMU, MMBench, MVBench, and V*Bench—demonstrate that GAM-Agent significantly improves performance across various VLM backbones. Notably, GAM-Agent boosts the accuracy of small-to-mid scale models (e.g., Qwen2.5-VL-7B, InternVL3-14B) by 5–6\%, and still enhances strong models like GPT-4o by up to 2–3\%. Our approach is modular, scalable, and generalizable, offering a path toward reliable and explainable multi-agent multimodal reasoning.

Video instance segmentation on mobile devices is an important yet very challenging edge AI problem. It mainly suffers from (1) heavy computation and memory costs for frame-by-frame pixel-level instance perception and (2) complicated heuristics for tracking objects. To address these issues, we present MobileInst, a lightweight and mobile-friendly framework for video instance segmentation on mobile devices. Firstly, MobileInst adopts a mobile vision transformer to extract multi-level semantic features and presents an efficient query-based dual-transformer instance decoder for mask kernels and a semantic-enhanced mask decoder to generate instance segmentation per frame. Secondly, MobileInst exploits simple yet effective kernel reuse and kernel association to track objects for video instance segmentation. Further, we propose temporal query passing to enhance the tracking ability for kernels. We conduct experiments on COCO and YouTube-VIS datasets to demonstrate the superiority of MobileInst and evaluate the inference latency on one single CPU core of the Snapdragon 778G Mobile Platform, without other methods of acceleration. On the COCO dataset, MobileInst achieves 31.2 mask AP and 433 ms on the mobile CPU, which reduces the latency by 50% compared to the previous SOTA. For video instance segmentation, MobileInst achieves 35.0 AP and 30.1 AP on YouTube-VIS 2019 & 2021.

3D Semantic Scene Completion (SSC) has emerged as a nascent and pivotal undertaking in autonomous driving, aiming to predict the voxel occupancy within volumetric scenes. However, prevailing methodologies primarily focus on voxel-wise feature aggregation, while neglecting instance semantics and scene context. In this paper, we present a novel paradigm termed Symphonies (Scene-from-Insts), that delves into the integration of instance queries to orchestrate 2D-to-3D reconstruction and 3D scene modeling. Leveraging our proposed Serial Instance-Propagated Attentions, Symphonies dynamically encodes instance-centric semantics, facilitating intricate interactions between the image and volumetric domains. Simultaneously, Symphonies fosters holistic scene comprehension by capturing context through the efficient fusion of instance queries, alleviating geometric ambiguities such as occlusion and perspective errors through contextual scene reasoning. Experimental results demonstrate that Symphonies achieves state-of-the-art performance on the chal-lenging SemanticKITTI and SSCBench-KITTI-360 benchmarks, yielding remarkable mIoU scores of 15.04 and 18.58, respectively. These results showcase the promising advancements of our paradigm. The code for our method is available at https://github.com/hustvl/Symphonies.

In this paper, a fault-tolerant finite control set model predictive current control (FCS-MPCC) method based on the extended voltage vector and adaptive control set (FCS-MPCC-EACS) is proposed, which improves the fault-tolerant control accuracy and stability of five-phase permanent magnet synchronous motor (F-PMSM) under two-phase open-circuit fault (OCF), and reduces the computational burden. Firstly, the extended voltage vector is proposed, and the optimal voltage vector and duty ratio are calculated by optimizing the cost function, thereby improving the fault-tolerant control accuracy and stability of the F-PMSM. Then, the deadbeat (DB) method is utilized to calculate the reference voltage vector for constructing the adaptive control set, which replaces the complete voltage vector enumeration process, thereby reducing the computational burden. Simulation test results show that compared with the FCS-MPCC method, FCS-MPCC with the extended voltage vector control set (FCS-MPCC-ECS) and FCS-MPCC-EACS methods improve the speed fault-tolerant control accuracy under non-adjacent two-phase OCF by 14.9% and 17.4%, respectively. At the same time, compared with the FCS-MPCC-ECS method, the calculation time of the FCS-MPCC-EACS method is shortened by 83.2%.

Due to low computational burden, high control bandwidth, excellent dynamic performance, and no overshoot, the deadbeat model predictive current control (DB-MPCC) method has great application potential in five-phase permanent magnet synchronous motors (F-PMSMs). However, the high dependence on the accuracy of the prediction model seriously affects the practical performance and application of the DB-MPCC method. This paper proposes an improved incremental model based DB-MPCC (IIM-DB-MPCC) method for F-PMSMs, which improves the discretization accuracy of the prediction model and effectively reduces the dependence of the DB-MPCC method on the accuracy of the prediction model. Firstly, the prediction model of the F-PMSM is discretized using the zero-order holding (ZOH) method, and the quantized control of the model discretization error is realized. Then, based on the improved discrete prediction model, the IIM-DB-MPCC method is proposed. Simulation test results show that, compared with the DB-MPCC method, the IIM-DB-MPCC method proposed in this paper can obtain higher prediction model accuracy than the Euler method at a low computational burden cost. Moreover, the static error caused by the mismatch of prediction model parameters is eliminated and the current and speed control accuracy of the F-PMSM is improved.