Ziyu Wan

DB-VIO: Dual-Branch Visual Inertial Odometry with Enhanced Visual-Inertial Representation

Ziyu Wan, Lin Zhao

Visual inertial odometry (VIO) is essential for accurate 6-DoF motion estimation in mobile robotic systems. Recent learning-based VIO methods have shown promising progress, but they often rely on unified visual--inertial representations and a single temporal model for full-pose estimation, limiting their ability …

DB-VIO: Dual-Branch Visual Inertial Odometry with Enhanced Visual-Inertial Representation

Ziyu Wan, Lin Zhao

Visual inertial odometry (VIO) is essential for accurate 6-DoF motion estimation in mobile robotic systems. Recent learning-based VIO methods have shown promising progress, but they often rely on unified visual--inertial representations and a single temporal model for full-pose estimation, limiting their ability …

VLM4D: Towards Spatiotemporal Awareness in Vision Language Models

Shijie Zhou, Alexander Vilesov, Xuehai He, Ziyu Wan et autres

Vision language models (VLMs) have shown remarkable capabilities in integrating linguistic and visual reasoning but remain fundamentally limited in understanding dynamic spatiotemporal interactions. Humans effortlessly track and reason about object movements, rotations, and perspective shifts-abilities essential for robust dynamic real-world understanding yet …

us, gb (code pays fourni par la source)

AvatarArtist: Open-Domain 4D Avatarization

Hongyu Liu, Xuan Wang, Ziyu Wan, Yue Ma et autres

This work focuses on open-domain 4D avatarization, with the purpose of creating a 4D avatar from a portrait image in an arbitrary style. We select parametric triplanes as the intermediate 4D representation, and propose a practical training paradigm that takes advantage of …

hk, cn (code pays fourni par la source)

AvatarArtist: Open-Domain 4D Avatarization

Hongyu Liu, Xuan Wang, Ziyu Wan, Yue Ma et autres

This work focuses on open-domain 4D avatarization, with the purpose of creating a 4D avatar from a portrait image in an arbitrary style. We select parametric triplanes as the intermediate 4D representation and propose a practical training paradigm that takes advantage of …

An Object is Worth 64×64 Pixels: Generating 3D Object via Image Diffusion

Xingguang Yan, Han-Hung Lee, Ziyu Wan, Anne Lynn S. Chang

We introduce a new approach for generating realistic 3D models with UV maps through a representation termed “Object Images.” This approach encapsulates surface geometry, appearance, and patch structures within a 64×64 pixel image, effectively converting complex 3D shapes into a more manageable …

ca, hk (code pays fourni par la source)

Language Games as the Pathway to Artificial Superhuman Intelligence

Ying Wen, Ziyu Wan, Zhang Shao

The evolution of large language models (LLMs) toward artificial superhuman intelligence (ASI) hinges on data reproduction, a cyclical process in which models generate, curate and retrain on novel data to refine capabilities. Current methods, however, risk getting stuck in a data reproduction …

HeadArtist-VL: Vision / Language Guided 3D Head Generation with Self Score Distillation

Hongyu Liu, Xuan Wang, Ziyu Wan, Yujun Shen et autres

We present HeadArtist-VL, a 3D head generation method that suits either vision or language input. With a landmark-guided ControlNet serving as a generative prior, we come up with an efficient pipeline that optimizes a parameterized 3D head model under the supervision of …

hk, cn (code pays fourni par la source)

ReMA: Learning to Meta-Think for LLMs with Multi-agent Reinforcement Learning

Ziyu Wan, Yunxiang Li, Xiaoyu Wen, Yan Song et autres

Recent research on Reasoning of Large Language Models (LLMs) has sought to further enhance their performance by integrating meta-thinking -- enabling models to monitor, evaluate, and control their reasoning processes for more adaptive and effective problem-solving. However, current single-agent work lacks a …

cn, ca, gb (code pays fourni par la source)

Natural Language Reinforcement Learning

Xidong Feng, Bo Liu, Haotian Fu, Ziyu Wan et autres

Artificial intelligence progresses towards the "Era of Experience," where agents are expected to learn from continuous, grounded interaction. We argue that traditional Reinforcement Learning (RL), which typically represents value as a scalar, can restrict agent's deep understanding of environments and hinders the …

RaFE: Generative Radiance Fields Restoration

Zhongkai Wu, Ziyu Wan, Jing Zhang, Jing Liao et autres

cn, hk (code pays fourni par la source)

Prediction of typhoon-induced disaster activities of transmission towers based on random forest algorithm

Fengrui Cheng, Ziyu Wan, Wanqing He, Jian Rong Qiu et autres