Xinbo Wang

Weilun Feng, Chuanguang Yang, Haotong Qin et al.

Diffusion transformers (DiT) have demonstrated exceptional performance in video generation. However, their large number of parameters and high computational complexity limit their deployment on edge devices. Quantization can reduce storage requirements and accelerate inference by lowering the bit-width of model parameters. Yet, existing quantization methods for image generation models do not generalize well to video generation tasks. We identify two primary challenges: the loss of information during quantization and the misalignment between optimization objectives and the unique requirements of video generation. To address these challenges, we present Q-VDiT, a quantization framework specifically designed for video DiT models. From the quantization perspective, we propose the Token-aware Quantization Estimator (TQE), which compensates for quantization errors in both the token and feature dimensions. From the optimization perspective, we introduce Temporal Maintenance Distillation (TMD), which preserves the spatiotemporal correlations between frames and enables the optimization of each frame with respect to the overall video context. Our W3A6 Q-VDiT achieves a scene consistency of 23.40, setting a new benchmark and outperforming current state-of-the-art quantization methods by 1.9$\times$. Code will be available at https://github.com/cantbebetter2/Q-VDiT.

Huining Yuan, Zelai Xu, Zheyue Tan et al.

Developing Large Language Models (LLMs) to cooperate and compete effectively within multi-agent systems is a critical step towards more advanced intelligence. While reinforcement learning (RL) has proven effective for enhancing reasoning in single-agent tasks, its extension to multi-turn, multi-agent scenarios remains underexplored due to the challenges of long-horizon credit assignment and agent-specific advantage estimation. To address these challenges, we introduce MARS, an end-to-end RL framework that incentivizes Multi-Agent Reasoning of LLMs through Self-play in both cooperative and competitive games. MARS features a turn-level advantage estimator that aligns learning signals with each interaction for credit assignment, and an agent-specific advantage normalization to stabilize multi-agent training. By learning with self-play across cooperative and competitive games, the MARS agent trained from Qwen3-4B develops strong strategic abilities that generalize to held-out games with up to 28.7% performance improvements. More importantly, the capability acquired through self-play generalizes beyond games, yielding consistent performance gains of multi-agent systems in reasoning benchmarks. When integrated into leading multi-agent systems, our MARS agent achieves significant performance gains of 10.0% on AIME and 12.5% on GPQA-Diamond. These results establish end-to-end RL training with self-play in strategic games as a powerful approach for developing generalizable multi-agent reasoning capabilities in LLMs. Our code and models are publicly available at https://github.com/thu-nics/MARS.

Xinbo Wang, Shian Jia, Ziyang Huang et al.

Modern operating system schedulers employ a single, static policy, which struggles to deliver optimal performance across the diverse and dynamic workloads of contemporary systems. This "one-policy-fits-all" approach leads to significant compromises in fairness, throughput, and latency, particularly with the rise of heterogeneous hardware and varied application architectures. This paper proposes a new paradigm: dynamically selecting the optimal policy from a portfolio of specialized schedulers rather than designing a single, monolithic one. We present the Adaptive Scheduling Agent (ASA), a lightweight framework that intelligently matches workloads to the most suitable "expert" scheduling policy at runtime. ASA's core is a novel, low-overhead offline/online approach. First, an offline process trains a universal, hardware-agnostic machine learning model to recognize abstract workload patterns from system behaviors. Second, at runtime, ASA continually processes the model's predictions using a time-weighted probability voting algorithm to identify the workload, then makes a scheduling decision by consulting a pre-configured, machine-specific mapping table to switch to the optimal scheduler via Linux's sched_ext framework. This decoupled architecture allows ASA to adapt to new hardware platforms rapidly without expensive retraining of the core recognition model. Our evaluation, based on a novel benchmark focused on user-experience metrics, demonstrates that ASA consistently outperforms the default Linux scheduler (EEVDF), achieving superior results in 86.4% of test scenarios. Furthermore, ASA's selections are near-optimal, ranking among the top three schedulers in 78.6% of all scenarios. This validates our approach as a practical path toward more intelligent, adaptive, and responsive operating system schedulers.

Xiangyuan Peng, Yu Wang, Miao Tang et al.

Reliable autonomous driving systems require accurate detection of traffic participants. To this end, multi-modal fusion has emerged as an effective strategy. In particular, 4D radar and LiDAR fusion methods based on multi-frame radar point clouds have demonstrated the effectiveness in bridging the point density gap. However, they often neglect radar point clouds' inter-frame misalignment caused by object movement during accumulation and do not fully exploit the object dynamic information from 4D radar. In this paper, we propose MoRAL, a motion-aware multi-frame 4D radar and LiDAR fusion framework for robust 3D object detection. First, a Motion-aware Radar Encoder (MRE) is designed to compensate for inter-frame radar misalignment from moving objects. Later, a Motion Attention Gated Fusion (MAGF) module integrate radar motion features to guide LiDAR features to focus on dynamic foreground objects. Extensive evaluations on the View-of-Delft (VoD) dataset demonstrate that MoRAL outperforms existing methods, achieving the highest mAP of 73.30% in the entire area and 88.68% in the driving corridor. Notably, our method also achieves the best AP of 69.67% for pedestrians in the entire area and 96.25% for cyclists in the driving corridor.