arXiv subjects

Jiacheng Fu

Publications and source records attributed to Jiacheng Fu.

3 recordsLinked to original sources

arXiv2026

4D-WAM: 4D Consistent World Modeling for Autonomous Driving

Emerging World-Action Models (WAMs) have demonstrated promising performance in autonomous driving by jointly modeling future driving scene evolution and trajectory planning. However, existing WAMs are typically trained with video data, which is only 2D projections of the underlying 4D driving scene. Consequently, WAMs fail to understand and capture the structure of 4D scenes and thus generate visually plausible yet 4D inconsistent future predictions that mislead downstream planning. To alleviate this issue, we present 4D-WAM, a model that leverages geometric foundation models for training-time supervision to enable 4D consistent world modeling. Specifically, we feed WAM-predicted future frames into a geometric foundation model, and use 4D-aware responses to define a 4D consistency loss. This loss encourages the model to understand, represent, and predict physically consistent 4D scenes during training, without additional inference cost. Moreover, we identify an early-decision phenomenon in WAMs and propose a decision-oriented timestep sampling strategy that emphasizes supervision at early, high-noise stages, where driving decisions are primarily formed. By propagating 4D supervision to this critical decision-formation phase, the proposed strategy further improves trajectory planning. Extensive experiments demonstrate that 4D-WAM effectively models 4D consistent scene evolution and achieves state-of-the-art performance on challenging NAVSIM-v1 and NAVSIM-v2 benchmarks.

cs.CV↗

arXiv2026

SUV: Future Scene Understanding as Video Generation for End-to-End Driving

End-to-end driving requires a coherent understanding of future scenes, yet existing methods model these scenes using task-specific heads and output formats, with limited scalability. Can video generation instead provide a shared predictor? We introduce SUV, a unified end-to-end driving framework that casts future Scene Understanding as Video generation using a pretrained video foundation model. SUV models future appearance, semantics, relative depth, and instance-level dynamics as video streams with a shared video expert, without stream-specific visual prediction heads. Through joint video-action attention, the action expert attends to the latent representations of all future streams and generates the ego trajectory. Experiments show that SUV directly predicts all four future streams, while controlled ablations show that structured future supervision and direct future-stream access yield higher trajectory planning scores. With only a single front camera and no candidate-trajectory selection, SUV outperforms a broad set of recent state-of-the-art methods on both NAVSIM-v2 splits, achieving 91.0 EPDMS on navtest and 36.9 on navhard. On the long-tail WOD-E2E benchmark, SUV achieves a competitive RFS of 7.94.

cs.CV↗

arXiv2023

Complex dynamics of knowledgeable monopoly models with gradient mechanisms

In this paper, we explore the dynamics of two monopoly models with knowledgeable players. The first model was initially introduced by Naimzada and Ricchiuti, while the second one is simplified from a famous monopoly introduced by Puu. We employ several tools based on symbolic computations to analyze the local stability and bifurcations of the two models. To the best of our knowledge, the complete stability conditions of the second model are obtained for the first time. We also investigate periodic solutions as well as their stability. Most importantly, we discover that the topological structure of the parameter space of the second model is much more complex than that of the first one. Specifically, in the first model, the parameter region for the stability of any periodic orbit with a fixed order constitutes a connected set. In the second model, however, the stability regions for the 3-cycle, 4-cycle, and 5-cycle orbits are disconnected sets formed by many disjoint portions. Furthermore, we find that the basins of the two stable equilibria in the second model are disconnected and also have complicated topological structures. In addition, the existence of chaos in the sense of Li-Yorke is rigorously proved by finding snapback repellers and 3-cycle orbits in the two models, respectively.

econ.TH↗